{"models":[{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.01,"canonical_slug":"anthropic/claude-haiku-5.5-20261007","catalog_created_at":"2026-10-07T18:31:23+00:00","catalog_status":"active","context_length":1000000,"eligible":true,"extra_pricing":{},"id":"anthropic/claude-haiku-5.5","input_price":0.1,"max_output_tokens":128000,"name":"Anthropic: Claude Haiku 5.5","new_model":true,"open_weights":false,"output_price":0.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthropic","provider_label":"anthropic","release_date":"2026-10-07","score":1424.6044605030932,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthropic/claude-haiku-5.5-20261007","version_policy":"independent_snapshot","evaluation_references":[{"benchmark":"Vals Index","benchmark_version":"2.2","checked_at":"2026-10-26T00:20:45.595508+00:00","date":"2026-10-24","effort":null,"identity_evidence":{"checked_at":"2026-10-26T00:20:45.595508+00:00","effort":null,"release_date":"2026-10-07","sha256":"06ea7e0887d58d20f5080ecbbb8b973f0471a6ee2eee78814bc7b478b1f97ad6","source_name":"Claude Haiku 5.5","url":"https://www.vals.ai/models/anthropic_claude-haiku-5-5"},"kind":"independent","model_url":"https://www.vals.ai/models/anthropic_claude-haiku-5-5","model_version":"anthropic/claude-haiku-5.5-20261007","note":"按经济领域权重汇总的任务成绩；原榜含部分其他模型使用备用模型的结果。推理档位为来源模型页默认值，单项测试可能不同，完整条件见原榜。","rank":12,"scope":"该版本综合榜全部已评测模型","score":58.71,"source_id":"vals","source_model_id":"Claude Haiku 5.5","source_url":"https://www.vals.ai/benchmarks/vals_index","total":27,"unit":"%","update_mode":"automatic","stale":false},{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":162.47,"ci_low":156.19,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-10-07"],"catalog_release_date":"2026-10-07","source_model_date":"2026-10-07","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"anthropic/claude-haiku-5.5-20261007","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":16,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":159.32,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-10-07","source_model_id":"Claude Haiku 5.5","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"high","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"high","mapping_type":"effort","mapping_note":"Arena 只收录了 high 档位；调用时需设置同一推理档位","source_model_id":"Claude Haiku 5.5 (High)","confidence_interval":[1416.3091666000496,1432.8997544061367],"vote_count":6080.0,"benchmark_stale":false,"listed_days_ago":19,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.1,"canonical_slug":"anthropic/claude-sonnet-5.5-20260928","catalog_created_at":"2026-09-28T18:04:46+00:00","catalog_status":"active","context_length":1000000,"eligible":true,"extra_pricing":{},"id":"anthropic/claude-sonnet-5.5","input_price":2.0,"max_output_tokens":128000,"name":"Anthropic: Claude Sonnet 5.5","new_model":true,"open_weights":false,"output_price":10.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthropic","provider_label":"anthropic","release_date":"2026-09-28","score":1477.512147734337,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthropic/claude-sonnet-5.5-20260928","version_policy":"independent_snapshot","evaluation_references":[{"benchmark":"Vals Index","benchmark_version":"2.2","checked_at":"2026-10-26T00:20:45.595508+00:00","date":"2026-10-24","effort":null,"identity_evidence":{"checked_at":"2026-10-26T00:20:45.595508+00:00","effort":null,"release_date":"2026-09-28","sha256":"d3f5e9342259a56564d40f8b2c059b4e3865d7d214b8c0523c2ccb83b409748d","source_name":"Claude Sonnet 5.5","url":"https://www.vals.ai/models/anthropic_claude-sonnet-5-5"},"kind":"independent","model_url":"https://www.vals.ai/models/anthropic_claude-sonnet-5-5","model_version":"anthropic/claude-sonnet-5.5-20260928","note":"按经济领域权重汇总的任务成绩；原榜含部分其他模型使用备用模型的结果。推理档位为来源模型页默认值，单项测试可能不同，完整条件见原榜。","rank":1,"scope":"该版本综合榜全部已评测模型","score":70.85,"source_id":"vals","source_model_id":"Claude Sonnet 5.5","source_url":"https://www.vals.ai/benchmarks/vals_index","total":27,"unit":"%","update_mode":"automatic","stale":false},{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":171.34,"ci_low":163.06,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-09-28"],"catalog_release_date":"2026-09-28","source_model_date":"2026-09-28","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"anthropic/claude-sonnet-5.5-20260928","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":6,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":166.18,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-09-28","source_model_id":"Claude Sonnet 5.5","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"xhigh","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"xhigh","mapping_type":"effort","mapping_note":"Arena 只收录了 xhigh 档位；调用时需设置同一推理档位","source_model_id":"Claude Sonnet 5.5 (xHigh)","confidence_interval":[1470.4899106790178,1484.5343847896563],"vote_count":8977.0,"benchmark_stale":false,"listed_days_ago":28,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.1,"canonical_slug":"google/gemini-4-argon-20261013","catalog_created_at":"2026-10-13T16:04:37+00:00","catalog_status":"active","context_length":1048576,"eligible":true,"extra_pricing":{},"id":"google/gemini-4-argon","input_price":2.0,"max_output_tokens":943718,"name":"Google: Gemini 4 Argon","new_model":true,"open_weights":false,"output_price":10.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2026-10-13","score":1517.1715740590357,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-4-argon-20261013","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"high","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"high","mapping_type":"effort","mapping_note":"Arena 只收录了 high 档位；调用时需设置同一推理档位","source_model_id":"Gemini 4 Argon (High)","confidence_interval":[1510.5556617539964,1523.787486364075],"vote_count":10083.0,"benchmark_stale":false,"listed_days_ago":13,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.2,"canonical_slug":"meta/muse-wonder-20261012","catalog_created_at":"2026-10-12T12:25:45+00:00","catalog_status":"active","context_length":1048576,"eligible":true,"extra_pricing":{},"id":"meta/muse-wonder","input_price":4.0,"max_output_tokens":943718,"name":"Meta: Muse Wonder","new_model":true,"open_weights":false,"output_price":20.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"meta","provider_label":"meta","release_date":"2026-10-12","score":1508.7101317781676,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"meta/muse-wonder-20261012","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":165.44,"ci_low":159.39,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-10-12"],"catalog_release_date":"2026-10-12","source_model_date":"2026-10-12","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"meta/muse-wonder-20261012","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":12,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":161.85,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-10-12","source_model_id":"Muse Wonder","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"max","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"max","mapping_type":"effort","mapping_note":"Arena 只收录了 max 档位；调用时需设置同一推理档位","source_model_id":"Muse Wonder (Max)","confidence_interval":[1500.5921270528231,1516.8281365035123],"vote_count":6001.0,"benchmark_stale":false,"listed_days_ago":14,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.1,"canonical_slug":"openai/gpt-6.1-sol-20260929","catalog_created_at":"2026-09-29T17:28:02+00:00","catalog_status":"active","context_length":1050000,"eligible":true,"extra_pricing":{},"id":"openai/gpt-6.1-sol","input_price":2.0,"max_output_tokens":128000,"name":"OpenAI: GPT-6.1 Sol","new_model":true,"open_weights":false,"output_price":10.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2026-09-29","score":1479.9433082251264,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-6.1-sol-20260929","version_policy":"independent_snapshot","evaluation_references":[{"benchmark":"Vals Index","benchmark_version":"2.2","checked_at":"2026-10-26T00:20:45.595508+00:00","date":"2026-10-24","effort":"max","identity_evidence":{"checked_at":"2026-10-26T00:20:45.595508+00:00","effort":"max","release_date":"2026-09-29","sha256":"821f4fc589e98e0eab9464c518b6bc28b954f5fef29665ed26d19b5b4a1ea7be","source_name":"GPT-6.1 Sol","url":"https://www.vals.ai/models/openai_gpt-6.1-sol"},"kind":"independent","model_url":"https://www.vals.ai/models/openai_gpt-6.1-sol","model_version":"openai/gpt-6.1-sol-20260929","note":"按经济领域权重汇总的任务成绩；原榜含部分其他模型使用备用模型的结果。推理档位为来源模型页默认值，单项测试可能不同，完整条件见原榜。","rank":7,"scope":"该版本综合榜全部已评测模型","score":63.94,"source_id":"vals","source_model_id":"GPT-6.1 Sol","source_url":"https://www.vals.ai/benchmarks/vals_index","total":27,"unit":"%","update_mode":"automatic","stale":false},{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":174.43,"ci_low":165.0,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-09-29"],"catalog_release_date":"2026-09-29","source_model_date":"2026-09-29","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-6.1-sol-20260929","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":3,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":168.37,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-09-29","source_model_id":"GPT-6.1 Sol","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"max","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"max","mapping_type":"effort","mapping_note":"Arena 只收录了 max 档位；调用时需设置同一推理档位","source_model_id":"GPT-6.1 Sol (Max)","confidence_interval":[1472.6302544137961,1487.2563620364567],"vote_count":8305.0,"benchmark_stale":false,"listed_days_ago":27,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":1.0,"canonical_slug":"anthropic/claude-5-fable-20260609","catalog_created_at":"2026-06-09T12:18:35+00:00","catalog_status":"active","context_length":1000000,"eligible":true,"extra_pricing":{},"id":"anthropic/claude-fable-5","input_price":10.0,"max_output_tokens":128000,"name":"Anthropic: Claude Fable 5","new_model":false,"open_weights":false,"output_price":50.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthropic","provider_label":"anthropic","release_date":"2026-06-09","score":1503.2756933519508,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthropic/claude-5-fable-20260609","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":168.18,"ci_low":160.53,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-06-09"],"catalog_release_date":"2026-06-09","source_model_date":"2026-06-09","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"anthropic/claude-5-fable-20260609","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":9,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":163.42,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-06-09","source_model_id":"Claude Fable 5","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"high","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"high","mapping_type":"effort","mapping_note":"Arena 只收录了 high 档位；调用时需设置同一推理档位","source_model_id":"Claude Fable 5 (High)","confidence_interval":[1499.0698371247897,1507.4815495791122],"vote_count":44253.0,"benchmark_stale":false,"listed_days_ago":139,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.25,"canonical_slug":"anthropic/claude-fable-5.1-20260831","catalog_created_at":"2026-09-01T18:03:58+00:00","catalog_status":"active","context_length":1000000,"eligible":true,"extra_pricing":{},"id":"anthropic/claude-fable-5.1","input_price":10.0,"max_output_tokens":128000,"name":"Anthropic: Claude Fable 5.1","new_model":false,"open_weights":false,"output_price":50.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthropic","provider_label":"anthropic","release_date":"2026-09-01","score":1503.1905093383516,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthropic/claude-fable-5.1-20260831","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"max","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"max","mapping_type":"effort","mapping_note":"Arena 只收录了 max 档位；调用时需设置同一推理档位","source_model_id":"Claude Fable 5.1 (Max)","confidence_interval":[1497.3185146345672,1509.0625040421362],"vote_count":16343.0,"benchmark_stale":false,"listed_days_ago":55,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.1,"canonical_slug":"anthropic/claude-4.5-haiku-20251001","catalog_created_at":"2025-10-15T17:00:38+00:00","catalog_status":"active","context_length":200000,"eligible":true,"extra_pricing":{},"id":"anthropic/claude-haiku-4.5","input_price":1.0,"max_output_tokens":64000,"name":"Anthropic: Claude Haiku 4.5","new_model":false,"open_weights":false,"output_price":5.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthropic","provider_label":"anthropic","release_date":"2025-10-15","score":1414.4702014364866,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthropic/claude-4.5-haiku-20251001","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"date","mapping_note":"Arena 条目为快照 20251001，OpenRouter 路由未标同一日期，按同一型号对应","source_model_id":"Claude Haiku 4.5 (20251001)","confidence_interval":[1412.0806291013428,1416.8597737716304],"vote_count":152393.0,"benchmark_stale":false,"listed_days_ago":376,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":1.5,"canonical_slug":"anthropic/claude-4.1-opus-20250805","catalog_created_at":"2025-08-05T16:33:11+00:00","catalog_status":"active","context_length":200000,"eligible":true,"extra_pricing":{},"id":"anthropic/claude-opus-4.1","input_price":15.0,"max_output_tokens":32000,"name":"Anthropic: Claude Opus 4.1","new_model":false,"open_weights":false,"output_price":75.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthropic","provider_label":"anthropic","release_date":"2025-08-05","score":1448.064495931572,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthropic/claude-4.1-opus-20250805","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":146.55,"ci_low":141.75,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-08-05"],"catalog_release_date":"2025-08-05","source_model_date":"2025-08-05","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"anthropic/claude-4.1-opus-20250805","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":95,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":144.07,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-08-05","source_model_id":"Claude Opus 4.1","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"date","mapping_note":"Arena 条目为快照 20250805，OpenRouter 路由未标同一日期，按同一型号对应","source_model_id":"Claude Opus 4.1 (20250805)","confidence_interval":[1445.094136189312,1451.0348556738322],"vote_count":77211.0,"benchmark_stale":false,"listed_days_ago":447,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.5,"canonical_slug":"anthropic/claude-4.6-opus-20260205","catalog_created_at":"2026-02-04T15:30:50+00:00","catalog_status":"active","context_length":1000000,"eligible":true,"extra_pricing":{},"id":"anthropic/claude-opus-4.6","input_price":5.0,"max_output_tokens":128000,"name":"Anthropic: Claude Opus 4.6","new_model":false,"open_weights":false,"output_price":25.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthropic","provider_label":"anthropic","release_date":"2026-02-05","score":1496.652551971709,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthropic/claude-4.6-opus-20260205","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":158.07,"ci_low":153.82,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-02-05"],"catalog_release_date":"2026-02-05","source_model_date":"2026-02-05","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"anthropic/claude-4.6-opus-20260205","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":33,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":155.66,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-02-05","source_model_id":"Claude Opus 4.6","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Claude Opus 4.6","confidence_interval":[1493.4472327856047,1499.8578711578134],"vote_count":88206.0,"benchmark_stale":false,"listed_days_ago":263,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.5,"canonical_slug":"anthropic/claude-4.7-opus-20260416","catalog_created_at":"2026-04-16T14:51:40+00:00","catalog_status":"active","context_length":1000000,"eligible":true,"extra_pricing":{},"id":"anthropic/claude-opus-4.7","input_price":5.0,"max_output_tokens":128000,"name":"Anthropic: Claude Opus 4.7","new_model":false,"open_weights":false,"output_price":25.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthropic","provider_label":"anthropic","release_date":"2026-04-16","score":1493.6753149588794,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthropic/claude-4.7-opus-20260416","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":159.19,"ci_low":154.91,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-04-16"],"catalog_release_date":"2026-04-16","source_model_date":"2026-04-16","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"anthropic/claude-4.7-opus-20260416","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":26,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":156.79,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-04-16","source_model_id":"Claude Opus 4.7","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Claude Opus 4.7","confidence_interval":[1489.915757952128,1497.4348719656307],"vote_count":66183.0,"benchmark_stale":false,"listed_days_ago":193,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.5,"canonical_slug":"anthropic/claude-4.8-opus-20260528","catalog_created_at":"2026-05-27T18:04:51+00:00","catalog_status":"active","context_length":1000000,"eligible":true,"extra_pricing":{},"id":"anthropic/claude-opus-4.8","input_price":5.0,"max_output_tokens":128000,"name":"Anthropic: Claude Opus 4.8","new_model":false,"open_weights":false,"output_price":25.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthropic","provider_label":"anthropic","release_date":"2026-05-28","score":1474.2127030951465,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthropic/claude-4.8-opus-20260528","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":161.95,"ci_low":156.89,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-05-28"],"catalog_release_date":"2026-05-28","source_model_date":"2026-05-28","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"anthropic/claude-4.8-opus-20260528","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":17,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":159.01,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-05-28","source_model_id":"Claude Opus 4.8","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Claude Opus 4.8","confidence_interval":[1470.3709758231098,1478.054430367183],"vote_count":65208.0,"benchmark_stale":false,"listed_days_ago":151,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.5,"canonical_slug":"anthropic/claude-opus-5-20260723","catalog_created_at":"2026-07-24T17:02:24+00:00","catalog_status":"active","context_length":1000000,"eligible":true,"extra_pricing":{},"id":"anthropic/claude-opus-5","input_price":5.0,"max_output_tokens":128000,"name":"Anthropic: Claude Opus 5","new_model":false,"open_weights":false,"output_price":25.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthropic","provider_label":"anthropic","release_date":"2026-07-24","score":1490.8728559881154,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthropic/claude-opus-5-20260723","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"max","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"max","mapping_type":"effort","mapping_note":"Arena 只收录了 max 档位；调用时需设置同一推理档位","source_model_id":"Claude Opus 5 (Max)","confidence_interval":[1486.1468781057313,1495.5988338704997],"vote_count":34526.0,"benchmark_stale":false,"listed_days_ago":94,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.2,"canonical_slug":"anthropic/claude-opus-5.5-20260921","catalog_created_at":"2026-09-22T16:32:12+00:00","catalog_status":"active","context_length":1000000,"eligible":true,"extra_pricing":{},"id":"anthropic/claude-opus-5.5","input_price":4.0,"max_output_tokens":128000,"name":"Anthropic: Claude Opus 5.5","new_model":false,"open_weights":false,"output_price":20.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthropic","provider_label":"anthropic","release_date":"2026-09-22","score":1508.7303522485386,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthropic/claude-opus-5.5-20260921","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"high","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"high","mapping_type":"effort","mapping_note":"Arena 只收录了 high 档位；调用时需设置同一推理档位","source_model_id":"Claude Opus 5.5 (High)","confidence_interval":[1501.935761820307,1515.52494267677],"vote_count":10322.0,"benchmark_stale":false,"listed_days_ago":34,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.3,"canonical_slug":"anthropic/claude-4.5-sonnet-20250929","catalog_created_at":"2025-09-29T16:01:16+00:00","catalog_status":"active","context_length":1000000,"eligible":true,"extra_pricing":{},"id":"anthropic/claude-sonnet-4.5","input_price":3.0,"max_output_tokens":64000,"name":"Anthropic: Claude Sonnet 4.5","new_model":false,"open_weights":false,"output_price":15.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthropic","provider_label":"anthropic","release_date":"2025-09-29","score":1454.9592302632007,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthropic/claude-4.5-sonnet-20250929","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":148.31,"ci_low":145.03,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-09-29"],"catalog_release_date":"2025-09-29","source_model_date":"2025-09-29","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"anthropic/claude-4.5-sonnet-20250929","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":75,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":146.74,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-09-29","source_model_id":"Claude Sonnet 4.5","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"date","mapping_note":"Arena 条目为快照 20250929，OpenRouter 路由未标同一日期，按同一型号对应","source_model_id":"Claude Sonnet 4.5 (20250929)","confidence_interval":[1452.0984491440067,1457.8200113823948],"vote_count":82650.0,"benchmark_stale":false,"listed_days_ago":392,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.3,"canonical_slug":"anthropic/claude-4.6-sonnet-20260217","catalog_created_at":"2026-02-17T15:43:10+00:00","catalog_status":"active","context_length":1000000,"eligible":true,"extra_pricing":{},"id":"anthropic/claude-sonnet-4.6","input_price":3.0,"max_output_tokens":128000,"name":"Anthropic: Claude Sonnet 4.6","new_model":false,"open_weights":false,"output_price":15.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthropic","provider_label":"anthropic","release_date":"2026-02-17","score":1472.082135371373,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthropic/claude-4.6-sonnet-20260217","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":155.06,"ci_low":150.2,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-02-17"],"catalog_release_date":"2026-02-17","source_model_date":"2026-02-17","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"anthropic/claude-4.6-sonnet-20260217","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":49,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":152.79,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-02-17","source_model_id":"Claude Sonnet 4.6","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Claude Sonnet 4.6","confidence_interval":[1468.5704058298538,1475.593864912892],"vote_count":70953.0,"benchmark_stale":false,"listed_days_ago":251,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.2,"canonical_slug":"anthropic/claude-sonnet-5-20260630","catalog_created_at":"2026-06-30T18:11:23+00:00","catalog_status":"active","context_length":1000000,"eligible":true,"extra_pricing":{},"id":"anthropic/claude-sonnet-5","input_price":2.0,"max_output_tokens":128000,"name":"Anthropic: Claude Sonnet 5","new_model":false,"open_weights":false,"output_price":10.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthropic","provider_label":"anthropic","release_date":"2026-06-30","score":1461.479847444079,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthropic/claude-sonnet-5-20260630","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":159.65,"ci_low":154.1,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-06-30"],"catalog_release_date":"2026-06-30","source_model_date":"2026-06-30","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"anthropic/claude-sonnet-5-20260630","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":25,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":156.88,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-06-30","source_model_id":"Claude Sonnet 5","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"high","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"high","mapping_type":"effort","mapping_note":"Arena 只收录了 high 档位；调用时需设置同一推理档位","source_model_id":"Claude Sonnet 5 (High)","confidence_interval":[1457.4158558614586,1465.5438390266995],"vote_count":51784.0,"benchmark_stale":false,"listed_days_ago":118,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"cohere/command-a-03-2025","catalog_created_at":"2025-03-13T19:32:22+00:00","catalog_status":"active","context_length":256000,"eligible":true,"extra_pricing":{},"id":"cohere/command-a","input_price":2.5,"max_output_tokens":8192,"name":"Cohere: Command A","new_model":false,"open_weights":true,"output_price":10.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"cohere","provider_label":"cohere","release_date":"2025-03-13","score":1353.700145035004,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"cohere/command-a-03-2025","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"canonical","mapping_note":"Arena 名称与 OpenRouter 规范 slug（含日期）一致","source_model_id":"Command A (03-2025)","confidence_interval":[1350.22866448664,1357.1716255833678],"vote_count":55758.0,"benchmark_stale":false,"listed_days_ago":592,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"cohere/command-r-08-2024","catalog_created_at":"2024-08-30T00:00:00+00:00","catalog_status":"active","context_length":128000,"eligible":true,"extra_pricing":{},"id":"cohere/command-r-08-2024","input_price":0.15,"max_output_tokens":4000,"name":"Cohere: Command R (08-2024)","new_model":false,"open_weights":true,"output_price":0.6,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"cohere","provider_label":"cohere","release_date":"2024-08-30","score":1250.4842103537844,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"cohere/command-r-08-2024","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Command R (08-2024)","confidence_interval":[1243.882260772712,1257.0861599348566],"vote_count":10140.0,"benchmark_stale":false,"listed_days_ago":787,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"deepseek/deepseek-v3.1-terminus","catalog_created_at":"2025-09-22T13:37:55+00:00","catalog_status":"active","context_length":163840,"eligible":true,"extra_pricing":{},"id":"deepseek/deepseek-v3.1-terminus","input_price":0.27,"max_output_tokens":147456,"name":"DeepSeek: DeepSeek V3.1 Terminus","new_model":false,"open_weights":true,"output_price":1.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"deepseek","provider_label":"deepseek","release_date":"2025-09-22","score":1414.902625016469,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"deepseek/deepseek-v3.1-terminus","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"DeepSeek-V3.1-Terminus","confidence_interval":[1405.1859725957215,1424.6192774372166],"vote_count":3608.0,"benchmark_stale":false,"listed_days_ago":399,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.028,"canonical_slug":"deepseek/deepseek-v3.2-20251201","catalog_created_at":"2025-12-01T13:10:42+00:00","catalog_status":"active","context_length":163840,"eligible":true,"extra_pricing":{},"id":"deepseek/deepseek-v3.2","input_price":0.28,"max_output_tokens":65536,"name":"DeepSeek: DeepSeek V3.2","new_model":false,"open_weights":true,"output_price":0.42,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"deepseek","provider_label":"deepseek","release_date":"2025-12-01","score":1424.7551354480886,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"deepseek/deepseek-v3.2-20251201","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":147.38,"ci_low":144.27,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-12-01"],"catalog_release_date":"2025-12-01","source_model_date":"2025-12-01","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"deepseek/deepseek-v3.2-20251201","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":82,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":146.09,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-12-01","source_model_id":"DeepSeek-V3.2","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"DeepSeek-V3.2","confidence_interval":[1421.2645712646563,1428.245699631521],"vote_count":48090.0,"benchmark_stale":false,"listed_days_ago":329,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"deepseek/deepseek-v3.2-exp","catalog_created_at":"2025-09-29T12:54:41+00:00","catalog_status":"active","context_length":163840,"eligible":true,"extra_pricing":{},"id":"deepseek/deepseek-v3.2-exp","input_price":0.27,"max_output_tokens":147456,"name":"DeepSeek: DeepSeek V3.2 Exp","new_model":false,"open_weights":true,"output_price":0.41,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"deepseek","provider_label":"deepseek","release_date":"2025-09-29","score":1421.7523608887582,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"deepseek/deepseek-v3.2-exp","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":146.81,"ci_low":142.45,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":[],"catalog_release_date":"2025-09-29","source_model_date":"2025-09-29","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"deepseek/deepseek-v3.2-exp","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":91,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":144.91,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-09-29","source_model_id":"DeepSeek-V3.2-Exp","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"DeepSeek-V3.2-Exp","confidence_interval":[1415.3727233979873,1428.131998379529],"vote_count":11997.0,"benchmark_stale":false,"listed_days_ago":392,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.0182,"canonical_slug":"deepseek/deepseek-v4-flash-20260423","catalog_created_at":"2026-04-24T03:17:46+00:00","catalog_status":"active","context_length":1048576,"eligible":true,"extra_pricing":{},"id":"deepseek/deepseek-v4-flash","input_price":0.091,"max_output_tokens":943718,"name":"DeepSeek: DeepSeek V4 Flash 0423","new_model":false,"open_weights":true,"output_price":0.182,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"deepseek","provider_label":"deepseek","release_date":"2026-04-24","score":1435.6906597797793,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"deepseek/deepseek-v4-flash-20260423","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"DeepSeek-V4-Flash","confidence_interval":[1431.6549812076933,1439.7263383518655],"vote_count":51982.0,"benchmark_stale":false,"listed_days_ago":185,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.0174,"canonical_slug":"deepseek/deepseek-v4-pro-20260423","catalog_created_at":"2026-04-24T03:17:59+00:00","catalog_status":"active","context_length":1048576,"eligible":true,"extra_pricing":{},"id":"deepseek/deepseek-v4-pro","input_price":0.2088,"max_output_tokens":384000,"name":"DeepSeek: DeepSeek V4 Pro 0423","new_model":false,"open_weights":true,"output_price":0.4176,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"deepseek","provider_label":"deepseek","release_date":"2026-04-24","score":1457.8787534988062,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"deepseek/deepseek-v4-pro-20260423","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"DeepSeek-V4-Pro","confidence_interval":[1453.9819517147716,1461.775555282841],"vote_count":57588.0,"benchmark_stale":false,"listed_days_ago":185,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.022,"canonical_slug":"deepseek/deepseek-v4-pro-20260813","catalog_created_at":"2026-08-12T15:42:44+00:00","catalog_status":"active","context_length":1048576,"eligible":true,"extra_pricing":{},"id":"deepseek/deepseek-v4-pro-0813","input_price":0.66,"max_output_tokens":393216,"name":"DeepSeek: DeepSeek V4 Pro 0813","new_model":false,"open_weights":true,"output_price":1.98,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"deepseek","provider_label":"deepseek","release_date":"2026-08-13","score":1464.5421674005563,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"official_release":{"api_alias":"deepseek-v4-pro","id":"deepseek/deepseek-v4-pro-0813","official_scores":[],"release_date":"2026-08-13","source_url":"https://api-docs.deepseek.com/updates/#deepseek-v4-pro-update","version":"DeepSeek-V4-Pro-0813","version_source_url":"https://api-docs.deepseek.com/quick_start/pricing/"},"official_source_stale":false,"model_version":"deepseek/deepseek-v4-pro-20260813","version_policy":"independent_snapshot","evaluation_references":[{"benchmark":"Vals Index","benchmark_version":"2.2","checked_at":"2026-10-26T00:20:45.595508+00:00","date":"2026-10-24","effort":"max","identity_evidence":{"checked_at":"2026-10-26T00:20:45.595508+00:00","effort":"max","release_date":"2026-08-13","sha256":"1f0cce438c4011133a3bc1d13a72373e8cfc921b0ab4905dfdcd65438031a2b2","source_name":"DeepSeek V4 Pro 0813","url":"https://www.vals.ai/models/deepseek_deepseek-v4-pro-0813"},"kind":"independent","model_url":"https://www.vals.ai/models/deepseek_deepseek-v4-pro-0813","model_version":"deepseek/deepseek-v4-pro-20260813","note":"按经济领域权重汇总的任务成绩；原榜含部分其他模型使用备用模型的结果。推理档位为来源模型页默认值，单项测试可能不同，完整条件见原榜。","rank":19,"scope":"该版本综合榜全部已评测模型","score":52.97,"source_id":"vals","source_model_id":"DeepSeek V4 Pro 0813","source_url":"https://www.vals.ai/benchmarks/vals_index","total":27,"unit":"%","update_mode":"automatic","stale":false},{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":158.04,"ci_low":153.74,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-08-13"],"catalog_release_date":"2026-08-13","source_model_date":"2026-08-13","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"deepseek/deepseek-v4-pro-20260813","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":34,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":155.64,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-08-13","source_model_id":"DeepSeek V4 Pro 0813","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"high","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"high","mapping_type":"effort","mapping_note":"Arena 只收录了 high 档位；调用时需设置同一推理档位","source_model_id":"DeepSeek-V4-Pro (High, 20260813)","confidence_interval":[1457.9318625831306,1471.1524722179822],"vote_count":9797.0,"benchmark_stale":false,"listed_days_ago":74,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.015,"canonical_slug":"deepseek/deepseek-v4.1-flash-20260910","catalog_created_at":"2026-09-10T06:21:25+00:00","catalog_status":"active","context_length":1048576,"eligible":true,"extra_pricing":{},"id":"deepseek/deepseek-v4.1-flash","input_price":0.02,"max_output_tokens":943718,"name":"DeepSeek: DeepSeek V4.1 Flash","new_model":false,"open_weights":true,"output_price":1.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"deepseek","provider_label":"deepseek","release_date":"2026-09-10","score":1476.0190205724746,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"deepseek/deepseek-v4.1-flash-20260910","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"max","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"max","mapping_type":"effort","mapping_note":"Arena 只收录了 max 档位；调用时需设置同一推理档位","source_model_id":"DeepSeek-V4.1-Flash-20260910 (Max)","confidence_interval":[1469.9430861693354,1482.094954975614],"vote_count":13750.0,"benchmark_stale":false,"listed_days_ago":46,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.35,"canonical_slug":"deepseek/deepseek-r1-0528","catalog_created_at":"2025-05-28T17:59:30+00:00","catalog_status":"active","context_length":163840,"eligible":true,"extra_pricing":{},"id":"deepseek/deepseek-r1-0528","input_price":0.5,"max_output_tokens":32768,"name":"DeepSeek: R1 0528","new_model":false,"open_weights":true,"output_price":2.15,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"deepseek","provider_label":"deepseek","release_date":"2025-05-28","score":1421.4038397193654,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"deepseek/deepseek-r1-0528","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"DeepSeek-R1 (0528)","confidence_interval":[1415.7258864463925,1427.0817929923383],"vote_count":18170.0,"benchmark_stale":false,"listed_days_ago":516,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.03,"canonical_slug":"google/gemini-2.5-flash","catalog_created_at":"2025-06-17T15:01:28+00:00","catalog_status":"active","context_length":1048576,"eligible":true,"extra_pricing":{},"id":"google/gemini-2.5-flash","input_price":0.3,"max_output_tokens":65535,"name":"Google: Gemini 2.5 Flash","new_model":false,"open_weights":false,"output_price":2.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2025-06-17","score":1409.2463611723172,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-2.5-flash","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Gemini 2.5 Flash","confidence_interval":[1406.8093192514968,1411.6834030931375],"vote_count":125462.0,"benchmark_stale":false,"listed_days_ago":496,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.125,"canonical_slug":"google/gemini-2.5-pro","catalog_created_at":"2025-06-17T14:12:24+00:00","catalog_status":"active","context_length":1048576,"eligible":true,"extra_pricing":{},"id":"google/gemini-2.5-pro","input_price":1.25,"max_output_tokens":65536,"name":"Google: Gemini 2.5 Pro","new_model":false,"open_weights":false,"output_price":10.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2025-06-17","score":1445.4765341377592,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-2.5-pro","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Gemini 2.5 Pro","confidence_interval":[1443.0153945960774,1447.937673679441],"vote_count":125316.0,"benchmark_stale":false,"listed_days_ago":496,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.025,"canonical_slug":"google/gemini-3.1-flash-lite-preview-20260303","catalog_created_at":"2026-03-03T04:37:53+00:00","catalog_status":"active","context_length":1048576,"eligible":true,"extra_pricing":{},"id":"google/gemini-3.1-flash-lite-preview","input_price":0.25,"max_output_tokens":65536,"name":"Google: Gemini 3.1 Flash Lite Preview","new_model":false,"open_weights":false,"output_price":1.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2026-03-03","score":1432.9648637006137,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-3.1-flash-lite-preview-20260303","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Gemini 3.1 Flash-Lite Preview","confidence_interval":[1429.3435518756482,1436.5861755255792],"vote_count":64209.0,"benchmark_stale":false,"listed_days_ago":237,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.2,"canonical_slug":"google/gemini-3.1-pro-preview-20260219","catalog_created_at":"2026-02-19T14:00:27+00:00","catalog_status":"active","context_length":1048576,"eligible":true,"extra_pricing":{},"id":"google/gemini-3.1-pro-preview","input_price":2.0,"max_output_tokens":65536,"name":"Google: Gemini 3.1 Pro Preview","new_model":false,"open_weights":false,"output_price":12.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2026-02-19","score":1487.018467203929,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-3.1-pro-preview-20260219","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Gemini 3.1 Pro Preview","confidence_interval":[1484.082580746095,1489.9543536617632],"vote_count":130819.0,"benchmark_stale":false,"listed_days_ago":249,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.15,"canonical_slug":"google/gemini-3.5-flash-20260519","catalog_created_at":"2026-05-19T12:30:00+00:00","catalog_status":"active","context_length":1048576,"eligible":true,"extra_pricing":{},"id":"google/gemini-3.5-flash","input_price":1.5,"max_output_tokens":65536,"name":"Google: Gemini 3.5 Flash","new_model":false,"open_weights":false,"output_price":9.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2026-05-19","score":1477.326500368638,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-3.5-flash-20260519","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":157.41,"ci_low":152.91,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-05-19"],"catalog_release_date":"2026-05-19","source_model_date":"2026-05-19","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"google/gemini-3.5-flash-20260519","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":39,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":154.88,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-05-19","source_model_id":"Gemini 3.5 Flash","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"high","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"high","mapping_type":"effort","mapping_note":"Arena 只收录了 high 档位；调用时需设置同一推理档位","source_model_id":"Gemini 3.5 Flash (High)","confidence_interval":[1473.43968706902,1481.213313668256],"vote_count":48504.0,"benchmark_stale":false,"listed_days_ago":160,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.03,"canonical_slug":"google/gemini-3.5-flash-lite-20260721","catalog_created_at":"2026-07-21T15:12:06+00:00","catalog_status":"active","context_length":1048576,"eligible":true,"extra_pricing":{},"id":"google/gemini-3.5-flash-lite","input_price":0.3,"max_output_tokens":65536,"name":"Google: Gemini 3.5 Flash Lite","new_model":false,"open_weights":false,"output_price":2.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2026-07-21","score":1455.1995844712576,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-3.5-flash-lite-20260721","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":146.26,"ci_low":142.41,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-07-21"],"catalog_release_date":"2026-07-21","source_model_date":"2026-07-21","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"google/gemini-3.5-flash-lite-20260721","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":91,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":144.91,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-07-21","source_model_id":"Gemini 3.5 Flash-Lite","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Gemini 3.5 Flash-Lite","confidence_interval":[1450.7313909934405,1459.6677779490744],"vote_count":35829.0,"benchmark_stale":false,"listed_days_ago":97,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.075,"canonical_slug":"google/gemini-3.6-flash-20260721","catalog_created_at":"2026-07-21T15:12:13+00:00","catalog_status":"active","context_length":1048576,"eligible":true,"extra_pricing":{},"id":"google/gemini-3.6-flash","input_price":0.75,"max_output_tokens":65536,"name":"Google: Gemini 3.6 Flash","new_model":false,"open_weights":false,"output_price":3.75,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2026-07-21","score":1483.0478951263665,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-3.6-flash-20260721","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":156.58,"ci_low":152.85,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-07-21"],"catalog_release_date":"2026-07-21","source_model_date":"2026-07-21","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"google/gemini-3.6-flash-20260721","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":41,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":154.53,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-07-21","source_model_id":"Gemini 3.6 Flash","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"high","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"high","mapping_type":"effort","mapping_note":"Arena 只收录了 high 档位；调用时需设置同一推理档位","source_model_id":"Gemini 3.6 Flash (High)","confidence_interval":[1478.628810933409,1487.4669793193239],"vote_count":36102.0,"benchmark_stale":false,"listed_days_ago":97,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.075,"canonical_slug":"google/gemini-3.7-flash-20260813","catalog_created_at":"2026-08-13T17:03:01+00:00","catalog_status":"active","context_length":1048576,"eligible":true,"extra_pricing":{},"id":"google/gemini-3.7-flash","input_price":0.75,"max_output_tokens":65536,"name":"Google: Gemini 3.7 Flash","new_model":false,"open_weights":false,"output_price":3.75,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2026-08-13","score":1489.8431217694936,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-3.7-flash-20260813","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":161.24,"ci_low":155.92,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-08-13"],"catalog_release_date":"2026-08-13","source_model_date":"2026-08-13","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"google/gemini-3.7-flash-20260813","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":18,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":157.97,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-08-13","source_model_id":"Gemini 3.7 Flash","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"high","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"high","mapping_type":"effort","mapping_note":"Arena 只收录了 high 档位；调用时需设置同一推理档位","source_model_id":"Gemini 3.7 Flash (High)","confidence_interval":[1485.1140034992363,1494.5722400397506],"vote_count":30501.0,"benchmark_stale":false,"listed_days_ago":74,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.0825,"canonical_slug":"google/gemini-3.8-flash-20260902","catalog_created_at":"2026-09-02T15:14:16+00:00","catalog_status":"active","context_length":1048576,"eligible":true,"extra_pricing":{},"id":"google/gemini-3.8-flash","input_price":0.825,"max_output_tokens":65536,"name":"Google: Gemini 3.8 Flash","new_model":false,"open_weights":false,"output_price":4.125,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2026-09-02","score":1495.7009350762085,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-3.8-flash-20260902","version_policy":"independent_snapshot","evaluation_references":[{"benchmark":"Vals Index","benchmark_version":"2.2","checked_at":"2026-10-26T00:20:45.595508+00:00","date":"2026-10-24","effort":"high","identity_evidence":{"checked_at":"2026-10-26T00:20:45.595508+00:00","effort":"high","release_date":"2026-09-02","sha256":"cb6d1f00c9e19f594bee14e69b6a72a9834895eb39d2d620567a5bfeda1fb2e5","source_name":"Gemini 3.8 Flash","url":"https://www.vals.ai/models/google_gemini-3.8-flash"},"kind":"independent","model_url":"https://www.vals.ai/models/google_gemini-3.8-flash","model_version":"google/gemini-3.8-flash-20260902","note":"按经济领域权重汇总的任务成绩；原榜含部分其他模型使用备用模型的结果。推理档位为来源模型页默认值，单项测试可能不同，完整条件见原榜。","rank":15,"scope":"该版本综合榜全部已评测模型","score":55.95,"source_id":"vals","source_model_id":"Gemini 3.8 Flash","source_url":"https://www.vals.ai/benchmarks/vals_index","total":27,"unit":"%","update_mode":"automatic","stale":false},{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":161.08,"ci_low":155.07,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-09-02"],"catalog_release_date":"2026-09-02","source_model_date":"2026-09-02","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"google/gemini-3.8-flash-20260902","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":20,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":157.44,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-09-02","source_model_id":"Gemini 3.8 Flash","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"high","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"high","mapping_type":"effort","mapping_note":"Arena 只收录了 high 档位；调用时需设置同一推理档位","source_model_id":"Gemini 3.8 Flash (High)","confidence_interval":[1491.2272735333424,1500.1745966190745],"vote_count":43997.0,"benchmark_stale":false,"listed_days_ago":54,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"google/gemma-2-27b-it","catalog_created_at":"2024-07-13T00:00:00+00:00","catalog_status":"active","context_length":8192,"eligible":true,"extra_pricing":{},"id":"google/gemma-2-27b-it","input_price":0.65,"max_output_tokens":2048,"name":"Google: Gemma 2 27B","new_model":false,"open_weights":true,"output_price":0.65,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2024-07-13","score":1289.4985297482065,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemma-2-27b-it","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Gemma 2 27B IT","confidence_interval":[1286.101591512512,1292.8954679839012],"vote_count":75754.0,"benchmark_stale":false,"listed_days_ago":835,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"google/gemma-3-12b-it","catalog_created_at":"2025-03-13T21:50:25+00:00","catalog_status":"active","context_length":131072,"eligible":true,"extra_pricing":{},"id":"google/gemma-3-12b-it","input_price":0.05,"max_output_tokens":16384,"name":"Google: Gemma 3 12B","new_model":false,"open_weights":true,"output_price":0.15,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2025-03-12","score":1341.7793041993204,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemma-3-12b-it","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":128.84,"ci_low":114.23,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":[],"catalog_release_date":"2025-03-12","source_model_date":"2025-03-12","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"google/gemma-3-12b-it","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":184,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":123.54,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-03-12","source_model_id":"Gemma 3 12B","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Gemma 3 12B IT","confidence_interval":[1332.268417098835,1351.2901912998057],"vote_count":3829.0,"benchmark_stale":false,"listed_days_ago":593,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.04,"canonical_slug":"google/gemma-3-27b-it","catalog_created_at":"2025-03-12T05:12:39+00:00","catalog_status":"active","context_length":131072,"eligible":true,"extra_pricing":{},"id":"google/gemma-3-27b-it","input_price":0.08,"max_output_tokens":117964,"name":"Google: Gemma 3 27B","new_model":false,"open_weights":true,"output_price":0.35,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2025-03-12","score":1365.3392812137033,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemma-3-27b-it","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":132.04,"ci_low":124.48,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":[],"catalog_release_date":"2025-03-12","source_model_date":"2025-03-12","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"google/gemma-3-27b-it","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":160,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":129.97,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-03-12","source_model_id":"Gemma 3 27B","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Gemma 3 27B IT","confidence_interval":[1361.6355456796018,1369.043016747805],"vote_count":46665.0,"benchmark_stale":false,"listed_days_ago":593,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.02,"canonical_slug":"google/gemma-3-4b-it","catalog_created_at":"2025-03-13T22:38:30+00:00","catalog_status":"active","context_length":131072,"eligible":true,"extra_pricing":{},"id":"google/gemma-3-4b-it","input_price":0.04,"max_output_tokens":117964,"name":"Google: Gemma 3 4B","new_model":false,"open_weights":true,"output_price":0.05,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2025-03-12","score":1303.2668621962205,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemma-3-4b-it","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":123.59,"ci_low":97.39,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":[],"catalog_release_date":"2025-03-12","source_model_date":"2025-03-12","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"google/gemma-3-4b-it","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":213,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":115.88,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-03-12","source_model_id":"Gemma 3 4B","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Gemma 3 4B IT","confidence_interval":[1293.950382992402,1312.583341400039],"vote_count":4171.0,"benchmark_stale":false,"listed_days_ago":593,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.015,"canonical_slug":"ibm-granite/granite-4.2-8b-20260831","catalog_created_at":"2026-08-31T20:06:20+00:00","catalog_status":"active","context_length":131072,"eligible":true,"extra_pricing":{},"id":"ibm-granite/granite-4.2-8b","input_price":0.06,"max_output_tokens":117964,"name":"IBM: Granite 4.2 8B","new_model":false,"open_weights":true,"output_price":0.25,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"ibm-granite","provider_label":"ibm-granite","release_date":"2026-08-25","score":1289.7857353610057,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"ibm-granite/granite-4.2-8b-20260831","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Granite 4.2 8B","confidence_interval":[1278.617007754103,1300.954462967908],"vote_count":3259.0,"benchmark_stale":false,"listed_days_ago":62,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.025,"canonical_slug":"inception/mercury-2-20260304","catalog_created_at":"2026-03-04T14:57:55+00:00","catalog_status":"active","context_length":128000,"eligible":true,"extra_pricing":{},"id":"inception/mercury-2","input_price":0.25,"max_output_tokens":50000,"name":"Inception: Mercury 2","new_model":false,"open_weights":false,"output_price":0.75,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"inception","provider_label":"inception","release_date":"2026-02-24","score":1342.957476420446,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"inception/mercury-2-20260304","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Mercury 2","confidence_interval":[1332.4315088041742,1353.483444036718],"vote_count":3193.0,"benchmark_stale":false,"listed_days_ago":244,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.15,"canonical_slug":"meta/muse-spark-1.1-20260709","catalog_created_at":"2026-07-16T15:29:01+00:00","catalog_status":"active","context_length":1048576,"eligible":true,"extra_pricing":{},"id":"meta/muse-spark-1.1","input_price":1.25,"max_output_tokens":943718,"name":"Meta: Muse Spark 1.1","new_model":false,"open_weights":false,"output_price":4.25,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"meta","provider_label":"meta","release_date":"2026-04-08","score":1489.8545257025703,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"meta/muse-spark-1.1-20260709","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Muse Spark 1.1","confidence_interval":[1485.7042832538693,1494.0047681512713],"vote_count":46121.0,"benchmark_stale":false,"listed_days_ago":201,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.15,"canonical_slug":"meta/muse-spark-1.2-20260805","catalog_created_at":"2026-08-05T19:48:07+00:00","catalog_status":"active","context_length":1048576,"eligible":true,"extra_pricing":{},"id":"meta/muse-spark-1.2","input_price":1.25,"max_output_tokens":943718,"name":"Meta: Muse Spark 1.2","new_model":false,"open_weights":false,"output_price":4.25,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"meta","provider_label":"meta","release_date":"2026-08-05","score":1491.3642050323658,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"meta/muse-spark-1.2-20260805","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":157.56,"ci_low":152.82,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-08-05"],"catalog_release_date":"2026-08-05","source_model_date":"2026-08-05","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"meta/muse-spark-1.2-20260805","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":38,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":155.05,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-08-05","source_model_id":"Muse Spark 1.2","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"xhigh","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"xhigh","mapping_type":"effort","mapping_note":"Arena 只收录了 xhigh 档位；调用时需设置同一推理档位","source_model_id":"Muse Spark 1.2 (xHigh)","confidence_interval":[1484.7893221450408,1497.9390879196906],"vote_count":10470.0,"benchmark_stale":false,"listed_days_ago":82,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.15,"canonical_slug":"meta/muse-spark-1.3-20260902","catalog_created_at":"2026-09-02T19:45:59+00:00","catalog_status":"active","context_length":1048576,"eligible":true,"extra_pricing":{},"id":"meta/muse-spark-1.3","input_price":1.25,"max_output_tokens":943718,"name":"Meta: Muse Spark 1.3","new_model":false,"open_weights":false,"output_price":4.25,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"meta","provider_label":"meta","release_date":"2026-09-02","score":1495.565274618029,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"meta/muse-spark-1.3-20260902","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":159.82,"ci_low":155.26,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-09-02"],"catalog_release_date":"2026-09-02","source_model_date":"2026-09-02","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"meta/muse-spark-1.3-20260902","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":23,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":157.21,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-09-02","source_model_id":"Muse Spark 1.3","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"max","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"max","mapping_type":"effort","mapping_note":"Arena 只收录了 max 档位；调用时需设置同一推理档位","source_model_id":"Muse Spark 1.3 (Max)","confidence_interval":[1490.231943725725,1500.8986055103333],"vote_count":20506.0,"benchmark_stale":false,"listed_days_ago":54,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"meta-llama/llama-3.1-70b-instruct","catalog_created_at":"2024-07-23T00:00:00+00:00","catalog_status":"active","context_length":131072,"eligible":true,"extra_pricing":{},"id":"meta-llama/llama-3.1-70b-instruct","input_price":0.4,"max_output_tokens":16384,"name":"Meta: Llama 3.1 70B Instruct","new_model":false,"open_weights":true,"output_price":0.4,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"meta-llama","provider_label":"meta-llama","release_date":"2024-07-23","score":1293.3674162220418,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"meta-llama/llama-3.1-70b-instruct","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Llama 3.1 70B Instruct","confidence_interval":[1289.61972105481,1297.1151113892738],"vote_count":55240.0,"benchmark_stale":false,"listed_days_ago":825,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.025,"canonical_slug":"meta-llama/llama-3.1-8b-instruct","catalog_created_at":"2024-07-23T00:00:00+00:00","catalog_status":"active","context_length":131072,"eligible":true,"extra_pricing":{},"id":"meta-llama/llama-3.1-8b-instruct","input_price":0.05,"max_output_tokens":117964,"name":"Meta: Llama 3.1 8B Instruct","new_model":false,"open_weights":true,"output_price":0.08,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"meta-llama","provider_label":"meta-llama","release_date":"2024-07-23","score":1211.2354356132837,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"meta-llama/llama-3.1-8b-instruct","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Llama 3.1 8B Instruct","confidence_interval":[1207.0736098741656,1215.3972613524018],"vote_count":49605.0,"benchmark_stale":false,"listed_days_ago":825,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"meta-llama/llama-3.2-1b-instruct","catalog_created_at":"2024-09-25T00:00:00+00:00","catalog_status":"active","context_length":60000,"eligible":true,"extra_pricing":{},"id":"meta-llama/llama-3.2-1b-instruct","input_price":0.027,"max_output_tokens":54000,"name":"Meta: Llama 3.2 1B Instruct","new_model":false,"open_weights":true,"output_price":0.201,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"meta-llama","provider_label":"meta-llama","release_date":"2024-09-25","score":1111.1587170172766,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"meta-llama/llama-3.2-1b-instruct","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Llama 3.2 1B Instruct","confidence_interval":[1103.2812607867136,1119.0361732478395],"vote_count":8045.0,"benchmark_stale":false,"listed_days_ago":761,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"meta-llama/llama-3.2-3b-instruct","catalog_created_at":"2024-09-25T00:00:00+00:00","catalog_status":"active","context_length":131072,"eligible":true,"extra_pricing":{},"id":"meta-llama/llama-3.2-3b-instruct","input_price":0.05,"max_output_tokens":117964,"name":"Meta: Llama 3.2 3B Instruct","new_model":false,"open_weights":true,"output_price":0.33,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"meta-llama","provider_label":"meta-llama","release_date":"2024-09-25","score":1166.8311875713239,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"meta-llama/llama-3.2-3b-instruct","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Llama 3.2 3B Instruct","confidence_interval":[1159.1304370883497,1174.531938054298],"vote_count":7936.0,"benchmark_stale":false,"listed_days_ago":761,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"meta-llama/llama-3.3-70b-instruct","catalog_created_at":"2024-12-06T17:28:57+00:00","catalog_status":"active","context_length":131072,"eligible":true,"extra_pricing":{},"id":"meta-llama/llama-3.3-70b-instruct","input_price":0.1,"max_output_tokens":16384,"name":"Meta: Llama 3.3 70B Instruct","new_model":false,"open_weights":true,"output_price":0.32,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"meta-llama","provider_label":"meta-llama","release_date":"2024-12-06","score":1318.0465054576455,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"meta-llama/llama-3.3-70b-instruct","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Llama 3.3 70B Instruct","confidence_interval":[1314.5418726235305,1321.5511382917605],"vote_count":54493.0,"benchmark_stale":false,"listed_days_ago":689,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.05,"canonical_slug":"meta-llama/llama-4-maverick-17b-128e-instruct","catalog_created_at":"2025-04-05T19:37:02+00:00","catalog_status":"active","context_length":524288,"eligible":true,"extra_pricing":{},"id":"meta-llama/llama-4-maverick","input_price":0.1875,"max_output_tokens":16384,"name":"Meta: Llama 4 Maverick","new_model":false,"open_weights":true,"output_price":0.6525,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"meta-llama","provider_label":"meta-llama","release_date":"2025-04-05","score":1326.9317197284624,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"meta-llama/llama-4-maverick-17b-128e-instruct","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"canonical","mapping_note":"Arena 名称与 OpenRouter 规范 slug（含日期）一致","source_model_id":"Llama 4 Maverick 17B 128E Instruct","confidence_interval":[1322.6524453277923,1331.2109941291326],"vote_count":39444.0,"benchmark_stale":false,"listed_days_ago":569,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"meta-llama/llama-4-scout-17b-16e-instruct","catalog_created_at":"2025-04-05T19:31:59+00:00","catalog_status":"active","context_length":1310720,"eligible":true,"extra_pricing":{},"id":"meta-llama/llama-4-scout","input_price":0.1,"max_output_tokens":16384,"name":"Meta: Llama 4 Scout","new_model":false,"open_weights":true,"output_price":0.3,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"meta-llama","provider_label":"meta-llama","release_date":"2025-04-05","score":1321.506640088552,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"meta-llama/llama-4-scout-17b-16e-instruct","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":131.19,"ci_low":125.13,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":[],"catalog_release_date":"2025-04-05","source_model_date":"2025-04-05","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"meta-llama/llama-4-scout-17b-16e-instruct","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":161,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":129.63,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-04-05","source_model_id":"Llama 4 Scout","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"canonical","mapping_note":"Arena 名称与 OpenRouter 规范 slug（含日期）一致","source_model_id":"Llama 4 Scout 17B 16E Instruct","confidence_interval":[1316.7688325173588,1326.2444476597452],"vote_count":29830.0,"benchmark_stale":false,"listed_days_ago":569,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"microsoft/phi-4","catalog_created_at":"2025-01-10T06:17:52+00:00","catalog_status":"active","context_length":16384,"eligible":true,"extra_pricing":{},"id":"microsoft/phi-4","input_price":0.07,"max_output_tokens":14745,"name":"Microsoft: Phi 4","new_model":false,"open_weights":true,"output_price":0.14,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"microsoft","provider_label":"microsoft","release_date":"2025-01-10","score":1256.298292775756,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"microsoft/phi-4","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Phi-4","confidence_interval":[1251.6869532062942,1260.9096323452177],"vote_count":24126.0,"benchmark_stale":false,"listed_days_ago":654,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"minimax/minimax-m1","catalog_created_at":"2025-06-17T22:46:54+00:00","catalog_status":"active","context_length":1000000,"eligible":true,"extra_pricing":{},"id":"minimax/minimax-m1","input_price":0.55,"max_output_tokens":40000,"name":"MiniMax: MiniMax M1","new_model":false,"open_weights":false,"output_price":2.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"minimax","provider_label":"minimax","release_date":"2025-06-17","score":1363.8260293283763,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"minimax/minimax-m1","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"MiniMax-M1","confidence_interval":[1359.5328255700288,1368.119233086724],"vote_count":34299.0,"benchmark_stale":false,"listed_days_ago":496,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"minimax/minimax-m2","catalog_created_at":"2025-10-23T20:41:33+00:00","catalog_status":"active","context_length":204800,"eligible":true,"extra_pricing":{},"id":"minimax/minimax-m2","input_price":0.255,"max_output_tokens":131072,"name":"MiniMax: MiniMax M2","new_model":false,"open_weights":true,"output_price":1.02,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"minimax","provider_label":"minimax","release_date":"2025-10-27","score":1343.5879554447472,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"minimax/minimax-m2","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"MiniMax-M2","confidence_interval":[1335.8483762303727,1351.327534659122],"vote_count":6855.0,"benchmark_stale":false,"listed_days_ago":364,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.03,"canonical_slug":"minimax/minimax-m2.5-20260211","catalog_created_at":"2026-02-12T15:01:42+00:00","catalog_status":"active","context_length":204800,"eligible":true,"extra_pricing":{},"id":"minimax/minimax-m2.5","input_price":0.3,"max_output_tokens":131072,"name":"MiniMax: MiniMax M2.5","new_model":false,"open_weights":true,"output_price":1.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"minimax","provider_label":"minimax","release_date":"2026-02-12","score":1390.7198475369528,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"minimax/minimax-m2.5-20260211","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"MiniMax-M2.5","confidence_interval":[1386.8319039025976,1394.607791171308],"vote_count":42791.0,"benchmark_stale":false,"listed_days_ago":256,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.06,"canonical_slug":"minimax/minimax-m2.7-20260318","catalog_created_at":"2026-03-18T12:24:57+00:00","catalog_status":"active","context_length":204800,"eligible":true,"extra_pricing":{},"id":"minimax/minimax-m2.7","input_price":0.3,"max_output_tokens":131072,"name":"MiniMax: MiniMax M2.7","new_model":false,"open_weights":true,"output_price":1.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"minimax","provider_label":"minimax","release_date":"2026-03-18","score":1414.9702489524907,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"minimax/minimax-m2.7-20260318","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":148.18,"ci_low":138.37,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-03-18"],"catalog_release_date":"2026-03-18","source_model_date":"2026-03-18","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"minimax/minimax-m2.7-20260318","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":87,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":145.78,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-03-18","source_model_id":"MiniMax-M2.7","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"MiniMax-M2.7","confidence_interval":[1411.6543855855514,1418.28611231943],"vote_count":87731.0,"benchmark_stale":false,"listed_days_ago":222,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.06,"canonical_slug":"minimax/minimax-m3-20260531","catalog_created_at":"2026-05-31T16:36:14+00:00","catalog_status":"active","context_length":1048576,"eligible":true,"extra_pricing":{},"id":"minimax/minimax-m3","input_price":0.3,"max_output_tokens":512000,"name":"MiniMax: MiniMax M3","new_model":false,"open_weights":true,"output_price":1.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"minimax","provider_label":"minimax","release_date":"2026-06-01","score":1439.812467462423,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"minimax/minimax-m3-20260531","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"MiniMax-M3","confidence_interval":[1436.0423709691675,1443.5825639556783],"vote_count":67030.0,"benchmark_stale":false,"listed_days_ago":147,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.04,"canonical_slug":"mistralai/mistral-medium-3","catalog_created_at":"2025-05-07T14:15:41+00:00","catalog_status":"active","context_length":131072,"eligible":true,"extra_pricing":{},"id":"mistralai/mistral-medium-3","input_price":0.4,"max_output_tokens":104857,"name":"Mistral: Mistral Medium 3","new_model":false,"open_weights":false,"output_price":2.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"mistralai","provider_label":"mistralai","release_date":"2025-05-07","score":1387.1923508045595,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"mistralai/mistral-medium-3","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":135.36,"ci_low":130.38,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":[],"catalog_release_date":"2025-05-07","source_model_date":"2025-05-07","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"mistralai/mistral-medium-3","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":149,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":134.02,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-05-07","source_model_id":"Mistral Medium 3","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"date","mapping_note":"Arena 条目为快照 2505，OpenRouter 路由未标同一日期，按同一型号对应","source_model_id":"Mistral Medium 3 (2505)","confidence_interval":[1382.4728284196701,1391.9118731894487],"vote_count":32776.0,"benchmark_stale":false,"listed_days_ago":537,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.04,"canonical_slug":"mistralai/mistral-medium-3.1","catalog_created_at":"2025-08-13T14:33:59+00:00","catalog_status":"active","context_length":131072,"eligible":true,"extra_pricing":{},"id":"mistralai/mistral-medium-3.1","input_price":0.4,"max_output_tokens":104857,"name":"Mistral: Mistral Medium 3.1","new_model":false,"open_weights":false,"output_price":2.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"mistralai","provider_label":"mistralai","release_date":"2025-08-12","score":1407.8222582661217,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"mistralai/mistral-medium-3.1","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"date","mapping_note":"Arena 条目为快照 2508，OpenRouter 路由未标同一日期，按同一型号对应","source_model_id":"Mistral Medium 3.1 (2508)","confidence_interval":[1405.2246665503553,1410.4198499818883],"vote_count":95285.0,"benchmark_stale":false,"listed_days_ago":440,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"mistralai/mistral-medium-3.5-20260430","catalog_created_at":"2026-04-30T17:33:59+00:00","catalog_status":"active","context_length":262144,"eligible":true,"extra_pricing":{},"id":"mistralai/mistral-medium-3-5","input_price":1.5,"max_output_tokens":209715,"name":"Mistral: Mistral Medium 3.5","new_model":false,"open_weights":false,"output_price":7.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"mistralai","provider_label":"mistralai","release_date":"2026-04-30","score":1426.4014493272962,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"mistralai/mistral-medium-3.5-20260430","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Mistral Medium 3.5","confidence_interval":[1420.0099724816469,1432.7929261729455],"vote_count":11791.0,"benchmark_stale":false,"listed_days_ago":179,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"mistralai/mistral-small-24b-instruct-2501","catalog_created_at":"2025-01-30T16:43:29+00:00","catalog_status":"active","context_length":32768,"eligible":true,"extra_pricing":{},"id":"mistralai/mistral-small-24b-instruct-2501","input_price":0.05,"max_output_tokens":16384,"name":"Mistral: Mistral Small 3","new_model":false,"open_weights":true,"output_price":0.08,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"mistralai","provider_label":"mistralai","release_date":"2025-01-30","score":1274.3900153105462,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"mistralai/mistral-small-24b-instruct-2501","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Mistral Small 24B Instruct (2501)","confidence_interval":[1268.4681454507877,1280.3118851703048],"vote_count":14681.0,"benchmark_stale":false,"listed_days_ago":634,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"mistralai/mistral-small-3.1-24b-instruct-2503","catalog_created_at":"2025-03-17T19:15:37+00:00","catalog_status":"active","context_length":128000,"eligible":true,"extra_pricing":{},"id":"mistralai/mistral-small-3.1-24b-instruct","input_price":0.351,"max_output_tokens":102400,"name":"Mistral: Mistral Small 3.1 24B","new_model":false,"open_weights":true,"output_price":0.555,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"mistralai","provider_label":"mistralai","release_date":"2025-03-17","score":1302.7308081388335,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"mistralai/mistral-small-3.1-24b-instruct-2503","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"canonical","mapping_note":"Arena 名称与 OpenRouter 规范 slug（含日期）一致","source_model_id":"Mistral Small 3.1 24B Instruct (2503)","confidence_interval":[1298.1858519788511,1307.2757642988158],"vote_count":32721.0,"benchmark_stale":false,"listed_days_ago":588,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.15,"canonical_slug":"moonshotai/kimi-k2.5-0127","catalog_created_at":"2026-01-27T04:11:16+00:00","catalog_status":"active","context_length":262144,"eligible":true,"extra_pricing":{},"id":"moonshotai/kimi-k2.5","input_price":0.5,"max_output_tokens":235929,"name":"MoonshotAI: Kimi K2.5","new_model":false,"open_weights":true,"output_price":2.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"moonshotai","provider_label":"moonshotai","release_date":"2026-01-27","score":1449.8685141924625,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"moonshotai/kimi-k2.5-0127","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":149.52,"ci_low":146.55,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":[],"catalog_release_date":"2026-01-27","source_model_date":"2026-01-27","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"moonshotai/kimi-k2.5-0127","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":69,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":148.07,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-01-27","source_model_id":"Kimi K2.5","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"thinking","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"thinking","mapping_type":"effort","mapping_note":"Arena 只收录了 思考模式 档位；调用时需设置同一推理档位","source_model_id":"Kimi K2.5 (Thinking)","confidence_interval":[1446.5908359165262,1453.1461924683988],"vote_count":74744.0,"benchmark_stale":false,"listed_days_ago":272,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.15,"canonical_slug":"moonshotai/kimi-k2.6-20260420","catalog_created_at":"2026-04-20T15:36:42+00:00","catalog_status":"active","context_length":262144,"eligible":true,"extra_pricing":{},"id":"moonshotai/kimi-k2.6","input_price":0.65,"max_output_tokens":235929,"name":"MoonshotAI: Kimi K2.6","new_model":false,"open_weights":true,"output_price":3.41,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"moonshotai","provider_label":"moonshotai","release_date":"2026-04-20","score":1460.6629424609475,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"moonshotai/kimi-k2.6-20260420","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":152.72,"ci_low":149.04,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-04-20"],"catalog_release_date":"2026-04-20","source_model_date":"2026-04-20","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"moonshotai/kimi-k2.6-20260420","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":55,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":150.94,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-04-20","source_model_id":"Kimi K2.6","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Kimi K2.6","confidence_interval":[1456.237181419436,1465.088703502459],"vote_count":40026.0,"benchmark_stale":false,"listed_days_ago":189,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.21,"canonical_slug":"moonshotai/kimi-k3-20260715","catalog_created_at":"2026-07-16T15:30:58+00:00","catalog_status":"active","context_length":1048576,"eligible":true,"extra_pricing":{},"id":"moonshotai/kimi-k3","input_price":0.239,"max_output_tokens":943718,"name":"MoonshotAI: Kimi K3","new_model":false,"open_weights":true,"output_price":10.8,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"moonshotai","provider_label":"moonshotai","release_date":"2026-07-16","score":1487.4144884315765,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"moonshotai/kimi-k3-20260715","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"max","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"max","mapping_type":"effort","mapping_note":"Arena 只收录了 max 档位；调用时需设置同一推理档位","source_model_id":"Kimi K3 (Max)","confidence_interval":[1482.9393674297442,1491.8896094334089],"vote_count":35379.0,"benchmark_stale":false,"listed_days_ago":102,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"openai/gpt-4-turbo","catalog_created_at":"2024-04-09T00:00:00+00:00","catalog_status":"active","context_length":128000,"eligible":true,"extra_pricing":{},"id":"openai/gpt-4-turbo","input_price":10.0,"max_output_tokens":4096,"name":"OpenAI: GPT-4 Turbo","new_model":false,"open_weights":false,"output_price":30.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2023-11-06","score":1324.5048119281448,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-4-turbo","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"date","mapping_note":"Arena 条目为快照 2024-04-09，OpenRouter 路由未标同一日期，按同一型号对应","source_model_id":"GPT-4 Turbo (2024-04-09)","confidence_interval":[1320.5783000619047,1328.4313237943848],"vote_count":98114.0,"benchmark_stale":false,"listed_days_ago":1085,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.5,"canonical_slug":"openai/gpt-4.1-2025-04-14","catalog_created_at":"2025-04-14T17:23:05+00:00","catalog_status":"active","context_length":1047576,"eligible":true,"extra_pricing":{},"id":"openai/gpt-4.1","input_price":2.0,"max_output_tokens":32768,"name":"OpenAI: GPT-4.1","new_model":false,"open_weights":false,"output_price":8.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2025-04-14","score":1414.624193219759,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-4.1-2025-04-14","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":138.34,"ci_low":133.57,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-04-14"],"catalog_release_date":"2025-04-14","source_model_date":"2025-04-14","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-4.1-2025-04-14","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":137,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":136.65,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-04-14","source_model_id":"GPT-4.1","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"canonical","mapping_note":"Arena 名称与 OpenRouter 规范 slug（含日期）一致","source_model_id":"GPT-4.1 (2025-04-14)","confidence_interval":[1410.9097746611822,1418.3386117783357],"vote_count":50302.0,"benchmark_stale":false,"listed_days_ago":560,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.1,"canonical_slug":"openai/gpt-4.1-mini-2025-04-14","catalog_created_at":"2025-04-14T17:23:01+00:00","catalog_status":"active","context_length":1047576,"eligible":true,"extra_pricing":{},"id":"openai/gpt-4.1-mini","input_price":0.4,"max_output_tokens":32768,"name":"OpenAI: GPT-4.1 Mini","new_model":false,"open_weights":false,"output_price":1.6,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2025-04-14","score":1382.8403919881755,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-4.1-mini-2025-04-14","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":136.49,"ci_low":130.97,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-04-14"],"catalog_release_date":"2025-04-14","source_model_date":"2025-04-14","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-4.1-mini-2025-04-14","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":145,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":134.89,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-04-14","source_model_id":"GPT-4.1 mini","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"canonical","mapping_note":"Arena 名称与 OpenRouter 规范 slug（含日期）一致","source_model_id":"GPT-4.1 mini (2025-04-14)","confidence_interval":[1378.5061095710444,1387.1746744053066],"vote_count":38701.0,"benchmark_stale":false,"listed_days_ago":560,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.025,"canonical_slug":"openai/gpt-4.1-nano-2025-04-14","catalog_created_at":"2025-04-14T17:22:49+00:00","catalog_status":"active","context_length":1047576,"eligible":true,"extra_pricing":{},"id":"openai/gpt-4.1-nano","input_price":0.1,"max_output_tokens":942818,"name":"OpenAI: GPT-4.1 Nano","new_model":false,"open_weights":false,"output_price":0.4,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2025-04-14","score":1322.2727133292392,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-4.1-nano-2025-04-14","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":131.82,"ci_low":123.33,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-04-14"],"catalog_release_date":"2025-04-14","source_model_date":"2025-04-14","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-4.1-nano-2025-04-14","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":162,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":129.54,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-04-14","source_model_id":"GPT-4.1 nano","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"canonical","mapping_note":"Arena 名称与 OpenRouter 规范 slug（含日期）一致","source_model_id":"GPT-4.1 nano (2025-04-14)","confidence_interval":[1314.5427892611335,1330.0026373973446],"vote_count":6103.0,"benchmark_stale":false,"listed_days_ago":560,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"openai/gpt-4o-2024-05-13","catalog_created_at":"2024-05-13T00:00:00+00:00","catalog_status":"active","context_length":128000,"eligible":true,"extra_pricing":{},"id":"openai/gpt-4o-2024-05-13","input_price":5.0,"max_output_tokens":4096,"name":"OpenAI: GPT-4o (2024-05-13)","new_model":false,"open_weights":false,"output_price":15.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2024-05-13","score":1346.1214691257276,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-4o-2024-05-13","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"GPT-4o (2024-05-13)","confidence_interval":[1342.6553597042018,1349.587578547253],"vote_count":112881.0,"benchmark_stale":false,"listed_days_ago":896,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":1.25,"canonical_slug":"openai/gpt-4o-2024-08-06","catalog_created_at":"2024-08-06T00:00:00+00:00","catalog_status":"active","context_length":128000,"eligible":true,"extra_pricing":{},"id":"openai/gpt-4o-2024-08-06","input_price":2.5,"max_output_tokens":16384,"name":"OpenAI: GPT-4o (2024-08-06)","new_model":false,"open_weights":false,"output_price":10.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2024-08-06","score":1335.5887734841322,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-4o-2024-08-06","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"GPT-4o (2024-08-06)","confidence_interval":[1331.3749390841137,1339.8026078841506],"vote_count":45499.0,"benchmark_stale":false,"listed_days_ago":811,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.075,"canonical_slug":"openai/gpt-4o-mini-2024-07-18","catalog_created_at":"2024-07-18T00:00:00+00:00","catalog_status":"active","context_length":128000,"eligible":true,"extra_pricing":{},"id":"openai/gpt-4o-mini-2024-07-18","input_price":0.15,"max_output_tokens":16384,"name":"OpenAI: GPT-4o-mini (2024-07-18)","new_model":false,"open_weights":false,"output_price":0.6,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2024-07-18","score":1317.5894865458797,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-4o-mini-2024-07-18","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"GPT-4o mini (2024-07-18)","confidence_interval":[1314.0393217718554,1321.1396513199043],"vote_count":68705.0,"benchmark_stale":false,"listed_days_ago":830,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.125,"canonical_slug":"openai/gpt-5-2025-08-07","catalog_created_at":"2025-08-07T17:23:33+00:00","catalog_status":"active","context_length":400000,"eligible":true,"extra_pricing":{},"id":"openai/gpt-5","input_price":1.25,"max_output_tokens":128000,"name":"OpenAI: GPT-5","new_model":false,"open_weights":false,"output_price":10.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2025-08-07","score":1434.809882843828,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5-2025-08-07","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":null,"ci_low":null,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-08-07"],"catalog_release_date":"2025-08-07","source_model_date":"2025-08-07","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-5-2025-08-07","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":61,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":150.0,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-08-07","source_model_id":"GPT-5","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"high","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"high","mapping_type":"effort","mapping_note":"Arena 只收录了 high 档位；调用时需设置同一推理档位","source_model_id":"GPT-5 (High)","confidence_interval":[1430.3043358937578,1439.3154297938984],"vote_count":31712.0,"benchmark_stale":false,"listed_days_ago":445,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.025,"canonical_slug":"openai/gpt-5-mini-2025-08-07","catalog_created_at":"2025-08-07T17:23:27+00:00","catalog_status":"active","context_length":400000,"eligible":true,"extra_pricing":{},"id":"openai/gpt-5-mini","input_price":0.25,"max_output_tokens":128000,"name":"OpenAI: GPT-5 Mini","new_model":false,"open_weights":false,"output_price":2.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2025-08-07","score":1389.8984222253432,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5-mini-2025-08-07","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":147.11,"ci_low":143.33,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-08-07"],"catalog_release_date":"2025-08-07","source_model_date":"2025-08-07","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-5-mini-2025-08-07","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":89,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":145.34,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-08-07","source_model_id":"GPT-5 mini","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"high","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"high","mapping_type":"effort","mapping_note":"Arena 只收录了 high 档位；调用时需设置同一推理档位","source_model_id":"GPT-5 mini (High)","confidence_interval":[1385.2500164044154,1394.5468280462708],"vote_count":26799.0,"benchmark_stale":false,"listed_days_ago":445,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.005,"canonical_slug":"openai/gpt-5-nano-2025-08-07","catalog_created_at":"2025-08-07T17:23:22+00:00","catalog_status":"active","context_length":400000,"eligible":true,"extra_pricing":{},"id":"openai/gpt-5-nano","input_price":0.05,"max_output_tokens":128000,"name":"OpenAI: GPT-5 Nano","new_model":false,"open_weights":false,"output_price":0.4,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2025-08-07","score":1337.7738388897887,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5-nano-2025-08-07","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":141.2,"ci_low":134.11,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-08-07"],"catalog_release_date":"2025-08-07","source_model_date":"2025-08-07","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-5-nano-2025-08-07","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":127,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":139.21,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-08-07","source_model_id":"GPT-5 nano","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"high","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"high","mapping_type":"effort","mapping_note":"Arena 只收录了 high 档位；调用时需设置同一推理档位","source_model_id":"GPT-5 nano (High)","confidence_interval":[1330.8548386396628,1344.692839139915],"vote_count":8134.0,"benchmark_stale":false,"listed_days_ago":445,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.125,"canonical_slug":"openai/gpt-5.1-20251113","catalog_created_at":"2025-11-13T18:58:25+00:00","catalog_status":"active","context_length":400000,"eligible":true,"extra_pricing":{},"id":"openai/gpt-5.1","input_price":1.25,"max_output_tokens":128000,"name":"OpenAI: GPT-5.1","new_model":false,"open_weights":false,"output_price":10.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2025-11-13","score":1438.5918000439824,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.1-20251113","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":151.74,"ci_low":148.36,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-11-13"],"catalog_release_date":"2025-11-13","source_model_date":"2025-11-13","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-5.1-20251113","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":63,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":149.76,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-11-13","source_model_id":"GPT-5.1","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"GPT-5.1","confidence_interval":[1435.0019616193135,1442.1816384686513],"vote_count":44419.0,"benchmark_stale":false,"listed_days_ago":347,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.175,"canonical_slug":"openai/gpt-5.2-20251211","catalog_created_at":"2025-12-10T18:02:55+00:00","catalog_status":"active","context_length":400000,"eligible":true,"extra_pricing":{},"id":"openai/gpt-5.2","input_price":1.75,"max_output_tokens":128000,"name":"OpenAI: GPT-5.2","new_model":false,"open_weights":false,"output_price":14.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2025-12-11","score":1435.4700807077963,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.2-20251211","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":155.42,"ci_low":152.05,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-12-11"],"catalog_release_date":"2025-12-11","source_model_date":"2025-12-11","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-5.2-20251211","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":46,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":153.66,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-12-11","source_model_id":"GPT-5.2","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"GPT-5.2","confidence_interval":[1432.2981229970123,1438.6420384185803],"vote_count":83828.0,"benchmark_stale":false,"listed_days_ago":319,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.25,"canonical_slug":"openai/gpt-5.4-20260305","catalog_created_at":"2026-03-05T18:12:32+00:00","catalog_status":"active","context_length":1050000,"eligible":true,"extra_pricing":{},"id":"openai/gpt-5.4","input_price":2.5,"max_output_tokens":128000,"name":"OpenAI: GPT-5.4","new_model":false,"open_weights":false,"output_price":15.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2026-03-05","score":1464.8231556982098,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.4-20260305","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":159.8,"ci_low":155.38,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-03-05"],"catalog_release_date":"2026-03-05","source_model_date":"2026-03-05","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-5.4-20260305","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":22,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":157.31,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-03-05","source_model_id":"GPT-5.4","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"GPT-5.4","confidence_interval":[1461.1373777534734,1468.5089336429462],"vote_count":67855.0,"benchmark_stale":false,"listed_days_ago":235,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.075,"canonical_slug":"openai/gpt-5.4-mini-20260317","catalog_created_at":"2026-03-17T11:49:38+00:00","catalog_status":"active","context_length":400000,"eligible":true,"extra_pricing":{},"id":"openai/gpt-5.4-mini","input_price":0.75,"max_output_tokens":128000,"name":"OpenAI: GPT-5.4 Mini","new_model":false,"open_weights":false,"output_price":4.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2026-03-17","score":1447.0944068597867,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.4-mini-20260317","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":150.47,"ci_low":146.87,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-03-17"],"catalog_release_date":"2026-03-17","source_model_date":"2026-03-17","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-5.4-mini-20260317","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":67,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":148.7,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-03-17","source_model_id":"GPT-5.4 Mini","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"high","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"high","mapping_type":"effort","mapping_note":"Arena 只收录了 high 档位；调用时需设置同一推理档位","source_model_id":"GPT-5.4 mini (High)","confidence_interval":[1443.3484391916516,1450.8403745279218],"vote_count":63451.0,"benchmark_stale":false,"listed_days_ago":223,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.02,"canonical_slug":"openai/gpt-5.4-nano-20260317","catalog_created_at":"2026-03-17T11:49:47+00:00","catalog_status":"active","context_length":400000,"eligible":true,"extra_pricing":{},"id":"openai/gpt-5.4-nano","input_price":0.2,"max_output_tokens":128000,"name":"OpenAI: GPT-5.4 Nano","new_model":false,"open_weights":false,"output_price":1.25,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2026-03-17","score":1401.3880769513587,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.4-nano-20260317","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":148.15,"ci_low":143.63,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-03-17"],"catalog_release_date":"2026-03-17","source_model_date":"2026-03-17","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-5.4-nano-20260317","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":83,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":146.03,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-03-17","source_model_id":"GPT-5.4 Nano","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"high","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"high","mapping_type":"effort","mapping_note":"Arena 只收录了 high 档位；调用时需设置同一推理档位","source_model_id":"GPT-5.4 nano (High)","confidence_interval":[1397.6102761091772,1405.16587779354],"vote_count":62375.0,"benchmark_stale":false,"listed_days_ago":223,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.5,"canonical_slug":"openai/gpt-5.5-20260423","catalog_created_at":"2026-04-24T17:31:33+00:00","catalog_status":"active","context_length":1050000,"eligible":true,"extra_pricing":{},"id":"openai/gpt-5.5","input_price":5.0,"max_output_tokens":128000,"name":"OpenAI: GPT-5.5","new_model":false,"open_weights":false,"output_price":30.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2026-04-23","score":1476.6876302100268,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.5-20260423","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":163.67,"ci_low":157.33,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-04-23"],"catalog_release_date":"2026-04-23","source_model_date":"2026-04-23","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-5.5-20260423","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":14,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":159.68,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-04-23","source_model_id":"GPT-5.5","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"GPT-5.5","confidence_interval":[1473.0021710693609,1480.373089350693],"vote_count":70998.0,"benchmark_stale":false,"listed_days_ago":186,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.02,"canonical_slug":"openai/gpt-5.6-luna-20260709","catalog_created_at":"2026-07-09T09:54:24+00:00","catalog_status":"active","context_length":1050000,"eligible":true,"extra_pricing":{},"id":"openai/gpt-5.6-luna","input_price":0.2,"max_output_tokens":128000,"name":"OpenAI: GPT-5.6 Luna","new_model":false,"open_weights":false,"output_price":1.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2026-07-09","score":1452.9827749897509,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.6-luna-20260709","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":159.26,"ci_low":154.19,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-07-09"],"catalog_release_date":"2026-07-09","source_model_date":"2026-07-09","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-5.6-luna-20260709","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":29,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":156.5,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-07-09","source_model_id":"GPT-5.6 Luna","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"xhigh","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"xhigh","mapping_type":"effort","mapping_note":"Arena 只收录了 xhigh 档位；调用时需设置同一推理档位","source_model_id":"GPT-5.6 Luna (xHigh)","confidence_interval":[1448.6902895900732,1457.2752603894285],"vote_count":41426.0,"benchmark_stale":false,"listed_days_ago":109,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.4,"canonical_slug":"openai/gpt-5.6-sol-20260709","catalog_created_at":"2026-07-09T09:54:10+00:00","catalog_status":"active","context_length":1050000,"eligible":true,"extra_pricing":{},"id":"openai/gpt-5.6-sol","input_price":4.0,"max_output_tokens":128000,"name":"OpenAI: GPT-5.6 Sol","new_model":false,"open_weights":false,"output_price":20.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2026-07-09","score":1485.1373541029157,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.6-sol-20260709","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":167.06,"ci_low":160.15,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-07-09"],"catalog_release_date":"2026-07-09","source_model_date":"2026-07-09","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-5.6-sol-20260709","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":11,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":162.79,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-07-09","source_model_id":"GPT-5.6 Sol","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"xhigh","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"xhigh","mapping_type":"effort","mapping_note":"Arena 只收录了 xhigh 档位；调用时需设置同一推理档位","source_model_id":"GPT-5.6 Sol (xHigh)","confidence_interval":[1480.879218216783,1489.3954899890484],"vote_count":42713.0,"benchmark_stale":false,"listed_days_ago":109,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.2,"canonical_slug":"openai/gpt-5.6-terra-20260709","catalog_created_at":"2026-07-09T09:54:17+00:00","catalog_status":"active","context_length":1050000,"eligible":true,"extra_pricing":{},"id":"openai/gpt-5.6-terra","input_price":2.0,"max_output_tokens":128000,"name":"OpenAI: GPT-5.6 Terra","new_model":false,"open_weights":false,"output_price":12.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2026-07-09","score":1467.2183652681626,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.6-terra-20260709","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":163.31,"ci_low":157.3,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-07-09"],"catalog_release_date":"2026-07-09","source_model_date":"2026-07-09","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-5.6-terra-20260709","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":13,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":160.09,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-07-09","source_model_id":"GPT-5.6 Terra","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"xhigh","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"xhigh","mapping_type":"effort","mapping_note":"Arena 只收录了 xhigh 档位；调用时需设置同一推理档位","source_model_id":"GPT-5.6 Terra (xHigh)","confidence_interval":[1463.0457120219335,1471.391018514392],"vote_count":46470.0,"benchmark_stale":false,"listed_days_ago":109,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":1.0,"canonical_slug":"openai/gpt-6-astra-20260903","catalog_created_at":"2026-09-04T20:13:58+00:00","catalog_status":"active","context_length":1050000,"eligible":true,"extra_pricing":{},"id":"openai/gpt-6-astra","input_price":10.0,"max_output_tokens":128000,"name":"OpenAI: GPT-6 Astra","new_model":false,"open_weights":false,"output_price":50.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2026-09-04","score":1476.479602855043,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-6-astra-20260903","version_policy":"independent_snapshot","evaluation_references":[{"benchmark":"Vals Index","benchmark_version":"2.2","checked_at":"2026-10-26T00:20:45.595508+00:00","date":"2026-10-24","effort":"max","identity_evidence":{"checked_at":"2026-10-26T00:20:45.595508+00:00","effort":"max","release_date":"2026-09-03","sha256":"80f9fa8969e5eea06ab19566d34e685f47fdee7ba991c2436282601c462468e4","source_name":"GPT-6 Astra","url":"https://www.vals.ai/models/openai_gpt-6-astra"},"kind":"independent","model_url":"https://www.vals.ai/models/openai_gpt-6-astra","model_version":"openai/gpt-6-astra-20260903","note":"按经济领域权重汇总的任务成绩；原榜含部分其他模型使用备用模型的结果。推理档位为来源模型页默认值，单项测试可能不同，完整条件见原榜。","rank":5,"scope":"该版本综合榜全部已评测模型","score":65.02,"source_id":"vals","source_model_id":"GPT-6 Astra","source_url":"https://www.vals.ai/benchmarks/vals_index","total":27,"unit":"%","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"max","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"max","mapping_type":"effort","mapping_note":"Arena 只收录了 max 档位；调用时需设置同一推理档位","source_model_id":"GPT-6 Astra (Max)","confidence_interval":[1470.4511154147863,1482.5080902952996],"vote_count":14076.0,"benchmark_stale":false,"listed_days_ago":52,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.01,"canonical_slug":"openai/gpt-6-luna-20260922","catalog_created_at":"2026-09-22T18:13:06+00:00","catalog_status":"active","context_length":1050000,"eligible":true,"extra_pricing":{},"id":"openai/gpt-6-luna","input_price":0.1,"max_output_tokens":128000,"name":"OpenAI: GPT-6 Luna","new_model":false,"open_weights":false,"output_price":0.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2026-09-22","score":1442.1952399830611,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-6-luna-20260922","version_policy":"independent_snapshot","evaluation_references":[{"benchmark":"Vals Index","benchmark_version":"2.2","checked_at":"2026-10-26T00:20:45.595508+00:00","date":"2026-10-24","effort":"max","identity_evidence":{"checked_at":"2026-10-26T00:20:45.595508+00:00","effort":"max","release_date":"2026-09-22","sha256":"e483f5e76f48dd7922c8433af499f9d2b7fb71933f159096d90911fd7915c589","source_name":"GPT-6 Luna","url":"https://www.vals.ai/models/openai_gpt-6-luna"},"kind":"independent","model_url":"https://www.vals.ai/models/openai_gpt-6-luna","model_version":"openai/gpt-6-luna-20260922","note":"按经济领域权重汇总的任务成绩；原榜含部分其他模型使用备用模型的结果。推理档位为来源模型页默认值，单项测试可能不同，完整条件见原榜。","rank":21,"scope":"该版本综合榜全部已评测模型","score":52.32,"source_id":"vals","source_model_id":"GPT-6 Luna","source_url":"https://www.vals.ai/benchmarks/vals_index","total":27,"unit":"%","update_mode":"automatic","stale":false},{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":159.43,"ci_low":154.32,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-09-22"],"catalog_release_date":"2026-09-22","source_model_date":"2026-09-22","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-6-luna-20260922","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":27,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":156.75,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-09-22","source_model_id":"GPT-6 Luna","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"max","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"max","mapping_type":"effort","mapping_note":"Arena 只收录了 max 档位；调用时需设置同一推理档位","source_model_id":"GPT-6 Luna (Max)","confidence_interval":[1437.8734866264294,1446.516993339693],"vote_count":72504.0,"benchmark_stale":false,"listed_days_ago":34,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.2,"canonical_slug":"openai/gpt-6-sol-20260922","catalog_created_at":"2026-09-22T18:12:55+00:00","catalog_status":"active","context_length":1050000,"eligible":true,"extra_pricing":{},"id":"openai/gpt-6-sol","input_price":2.0,"max_output_tokens":128000,"name":"OpenAI: GPT-6 Sol","new_model":false,"open_weights":false,"output_price":10.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2026-09-22","score":1459.032251281731,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-6-sol-20260922","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":167.44,"ci_low":160.92,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-09-22"],"catalog_release_date":"2026-09-22","source_model_date":"2026-09-22","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-6-sol-20260922","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":8,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":163.73,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-09-22","source_model_id":"GPT-6 Sol","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"max","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"max","mapping_type":"effort","mapping_note":"Arena 只收录了 max 档位；调用时需设置同一推理档位","source_model_id":"GPT-6 Sol (Max)","confidence_interval":[1453.0120922438434,1465.052410319619],"vote_count":14707.0,"benchmark_stale":false,"listed_days_ago":34,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.075,"canonical_slug":"openai/gpt-oss-120b","catalog_created_at":"2025-08-05T17:17:11+00:00","catalog_status":"active","context_length":131072,"eligible":true,"extra_pricing":{},"id":"openai/gpt-oss-120b","input_price":0.15,"max_output_tokens":65536,"name":"OpenAI: gpt-oss-120b","new_model":false,"open_weights":true,"output_price":0.6,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2025-08-05","score":1351.8138449598982,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-oss-120b","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":142.12,"ci_low":135.57,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":[],"catalog_release_date":"2025-08-05","source_model_date":"2025-08-05","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-oss-120b","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":123,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":139.79,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-08-05","source_model_id":"gpt-oss-120b","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"gpt-oss-120b","confidence_interval":[1347.4489128740222,1356.178777045774],"vote_count":30207.0,"benchmark_stale":false,"listed_days_ago":447,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.009,"canonical_slug":"openai/gpt-oss-20b","catalog_created_at":"2025-08-05T17:17:09+00:00","catalog_status":"active","context_length":131072,"eligible":true,"extra_pricing":{},"id":"openai/gpt-oss-20b","input_price":0.018,"max_output_tokens":32768,"name":"OpenAI: gpt-oss-20b","new_model":false,"open_weights":true,"output_price":0.09,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2025-08-05","score":1317.4278677948328,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-oss-20b","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":139.58,"ci_low":132.2,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":[],"catalog_release_date":"2025-08-05","source_model_date":"2025-08-05","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-oss-20b","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":134,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":137.7,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-08-05","source_model_id":"gpt-oss-20b","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"gpt-oss-20b","confidence_interval":[1311.0129735762032,1323.842762013462],"vote_count":10417.0,"benchmark_stale":false,"listed_days_ago":447,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":7.5,"canonical_slug":"openai/o1-2024-12-17","catalog_created_at":"2024-12-17T18:26:39+00:00","catalog_status":"active","context_length":200000,"eligible":true,"extra_pricing":{},"id":"openai/o1","input_price":15.0,"max_output_tokens":100000,"name":"OpenAI: o1","new_model":false,"open_weights":false,"output_price":60.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2024-12-05","score":1402.2739743374746,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/o1-2024-12-17","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"canonical","mapping_note":"Arena 名称与 OpenRouter 规范 slug（含日期）一致","source_model_id":"o1 (2024-12-17)","confidence_interval":[1397.7976353657264,1406.7503133092227],"vote_count":27807.0,"benchmark_stale":false,"listed_days_ago":690,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.5,"canonical_slug":"openai/o3-2025-04-16","catalog_created_at":"2025-04-16T17:10:57+00:00","catalog_status":"active","context_length":200000,"eligible":true,"extra_pricing":{},"id":"openai/o3","input_price":2.0,"max_output_tokens":100000,"name":"OpenAI: o3","new_model":false,"open_weights":false,"output_price":8.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2025-04-16","score":1431.7526724620568,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/o3-2025-04-16","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":148.52,"ci_low":144.83,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-04-16"],"catalog_release_date":"2025-04-16","source_model_date":"2025-04-16","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/o3-2025-04-16","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":76,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":146.73,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-04-16","source_model_id":"o3","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"canonical","mapping_note":"Arena 名称与 OpenRouter 规范 slug（含日期）一致","source_model_id":"o3 (2025-04-16)","confidence_interval":[1428.1217529728374,1435.383591951276],"vote_count":58894.0,"benchmark_stale":false,"listed_days_ago":558,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.55,"canonical_slug":"openai/o3-mini-2025-01-31","catalog_created_at":"2025-01-31T19:28:41+00:00","catalog_status":"active","context_length":200000,"eligible":true,"extra_pricing":{},"id":"openai/o3-mini","input_price":1.1,"max_output_tokens":100000,"name":"OpenAI: o3 Mini","new_model":false,"open_weights":false,"output_price":4.4,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2024-12-20","score":1348.0581683771238,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/o3-mini-2025-01-31","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"o3-mini","confidence_interval":[1344.502666381458,1351.6136703727898],"vote_count":56742.0,"benchmark_stale":false,"listed_days_ago":675,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.55,"canonical_slug":"openai/o3-mini-high-2025-01-31","catalog_created_at":"2025-02-12T15:03:31+00:00","catalog_status":"active","context_length":200000,"eligible":true,"extra_pricing":{},"id":"openai/o3-mini-high","input_price":1.1,"max_output_tokens":100000,"name":"OpenAI: o3 Mini High","new_model":false,"open_weights":false,"output_price":4.4,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2025-02-12","score":1363.5758400668938,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/o3-mini-high-2025-01-31","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"o3-mini (High)","confidence_interval":[1358.308223701599,1368.8434564321883],"vote_count":18589.0,"benchmark_stale":false,"listed_days_ago":621,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.275,"canonical_slug":"openai/o4-mini-2025-04-16","catalog_created_at":"2025-04-16T16:29:02+00:00","catalog_status":"active","context_length":200000,"eligible":true,"extra_pricing":{},"id":"openai/o4-mini","input_price":1.1,"max_output_tokens":100000,"name":"OpenAI: o4 Mini","new_model":false,"open_weights":false,"output_price":4.4,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2025-04-16","score":1390.9289036816638,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/o4-mini-2025-04-16","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":147.21,"ci_low":142.9,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-04-16"],"catalog_release_date":"2025-04-16","source_model_date":"2025-04-16","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/o4-mini-2025-04-16","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":88,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":145.46,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-04-16","source_model_id":"o4-mini","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"canonical","mapping_note":"Arena 名称与 OpenRouter 规范 slug（含日期）一致","source_model_id":"o4-mini (2025-04-16)","confidence_interval":[1386.9366514667743,1394.9211558965533],"vote_count":44851.0,"benchmark_stale":false,"listed_days_ago":558,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.052,"canonical_slug":"qwen/qwen-plus-2025-01-25","catalog_created_at":"2025-02-01T11:37:20+00:00","catalog_status":"active","context_length":1000000,"eligible":true,"extra_pricing":{},"id":"qwen/qwen-plus","input_price":0.26,"max_output_tokens":32768,"name":"Qwen: Qwen-Plus","new_model":false,"open_weights":false,"output_price":0.78,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2024-01-25","score":1345.8872688693154,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen-plus-2025-01-25","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"date","mapping_note":"Arena 条目为快照 0125，OpenRouter 路由未标同一日期，按同一型号对应","source_model_id":"Qwen-Plus (0125)","confidence_interval":[1337.5561060794819,1354.2184316591488],"vote_count":5819.0,"benchmark_stale":false,"listed_days_ago":1005,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"qwen/qwen3-235b-a22b-07-25","catalog_created_at":"2025-07-21T17:39:15+00:00","catalog_status":"active","context_length":262144,"eligible":true,"extra_pricing":{},"id":"qwen/qwen3-235b-a22b-2507","input_price":0.09,"max_output_tokens":16384,"name":"Qwen: Qwen3 235B A22B Instruct 2507","new_model":false,"open_weights":true,"output_price":0.55,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2025-07-21","score":1374.858434288959,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3-235b-a22b-07-25","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"date","mapping_note":"OpenRouter 路由 qwen/qwen3-235b-a22b-2507 带版本日期，Arena 条目未标日期，按同一型号对应","source_model_id":"Qwen3-235B-A22B","confidence_interval":[1370.1487217518734,1379.5681468260443],"vote_count":25949.0,"benchmark_stale":false,"listed_days_ago":462,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"qwen/qwen3-235b-a22b-thinking-2507","catalog_created_at":"2025-07-25T13:19:17+00:00","catalog_status":"active","context_length":128000,"eligible":true,"extra_pricing":{},"id":"qwen/qwen3-235b-a22b-thinking-2507","input_price":0.45,"max_output_tokens":16384,"name":"Qwen: Qwen3 235B A22B Thinking 2507","new_model":false,"open_weights":true,"output_price":3.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2025-07-25","score":1399.8259689620845,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3-235b-a22b-thinking-2507","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Qwen3-235B-A22B-Thinking (2507)","confidence_interval":[1393.2523353951963,1406.3996025289728],"vote_count":8821.0,"benchmark_stale":false,"listed_days_ago":458,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"qwen/qwen3-30b-a3b-04-28","catalog_created_at":"2025-04-28T22:16:44+00:00","catalog_status":"active","context_length":40960,"eligible":true,"extra_pricing":{},"id":"qwen/qwen3-30b-a3b","input_price":0.12,"max_output_tokens":16384,"name":"Qwen: Qwen3 30B A3B","new_model":false,"open_weights":true,"output_price":0.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2025-04-28","score":1326.65359092986,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3-30b-a3b-04-28","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Qwen3-30B-A3B","confidence_interval":[1321.9045830691718,1331.4025987905482],"vote_count":26155.0,"benchmark_stale":false,"listed_days_ago":546,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"qwen/qwen3-30b-a3b-instruct-2507","catalog_created_at":"2025-07-29T16:36:05+00:00","catalog_status":"active","context_length":262144,"eligible":true,"extra_pricing":{},"id":"qwen/qwen3-30b-a3b-instruct-2507","input_price":0.09,"max_output_tokens":235929,"name":"Qwen: Qwen3 30B A3B Instruct 2507","new_model":false,"open_weights":true,"output_price":0.3,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2025-07-28","score":1382.245938854449,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3-30b-a3b-instruct-2507","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Qwen3-30B-A3B-Instruct (2507)","confidence_interval":[1377.339521635446,1387.1523560734515],"vote_count":23306.0,"benchmark_stale":false,"listed_days_ago":455,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"qwen/qwen3-32b-04-28","catalog_created_at":"2025-04-28T21:32:25+00:00","catalog_status":"active","context_length":131072,"eligible":true,"extra_pricing":{},"id":"qwen/qwen3-32b","input_price":0.08,"max_output_tokens":16384,"name":"Qwen: Qwen3 32B","new_model":false,"open_weights":true,"output_price":0.28,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2025-04-28","score":1346.8316292111117,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3-32b-04-28","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Qwen3-32B","confidence_interval":[1337.3749832163064,1356.288275205917],"vote_count":3926.0,"benchmark_stale":false,"listed_days_ago":546,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"qwen/qwen3-next-80b-a3b-instruct-2509","catalog_created_at":"2025-09-11T17:36:53+00:00","catalog_status":"active","context_length":262144,"eligible":true,"extra_pricing":{},"id":"qwen/qwen3-next-80b-a3b-instruct","input_price":0.09,"max_output_tokens":16384,"name":"Qwen: Qwen3 Next 80B A3B Instruct","new_model":false,"open_weights":true,"output_price":1.1,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2025-09","score":1398.502742753379,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3-next-80b-a3b-instruct-2509","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Qwen3-Next-80B-A3B-Instruct","confidence_interval":[1393.7286114173926,1403.2768740893653],"vote_count":22755.0,"benchmark_stale":false,"listed_days_ago":null,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"qwen/qwen3-next-80b-a3b-thinking-2509","catalog_created_at":"2025-09-11T17:38:04+00:00","catalog_status":"active","context_length":262144,"eligible":true,"extra_pricing":{},"id":"qwen/qwen3-next-80b-a3b-thinking","input_price":0.15,"max_output_tokens":235929,"name":"Qwen: Qwen3 Next 80B A3B Thinking","new_model":false,"open_weights":true,"output_price":1.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2025-09","score":1369.2691020431268,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3-next-80b-a3b-thinking-2509","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Qwen3-Next-80B-A3B-Thinking","confidence_interval":[1363.4074369928176,1375.1307670934361],"vote_count":13473.0,"benchmark_stale":false,"listed_days_ago":null,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.1,"canonical_slug":"qwen/qwen3-vl-235b-a22b-instruct","catalog_created_at":"2025-09-23T23:04:47+00:00","catalog_status":"active","context_length":262144,"eligible":true,"extra_pricing":{},"id":"qwen/qwen3-vl-235b-a22b-instruct","input_price":0.21,"max_output_tokens":32768,"name":"Qwen: Qwen3 VL 235B A22B Instruct","new_model":false,"open_weights":true,"output_price":1.9,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2025-09-23","score":1413.2501370308555,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3-vl-235b-a22b-instruct","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Qwen3-VL-235B-A22B-Instruct","confidence_interval":[1406.7235951317107,1419.7766789300003],"vote_count":11312.0,"benchmark_stale":false,"listed_days_ago":398,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.3,"canonical_slug":"qwen/qwen3.5-397b-a17b-20260216","catalog_created_at":"2026-02-16T06:23:38+00:00","catalog_status":"active","context_length":262144,"eligible":true,"extra_pricing":{},"id":"qwen/qwen3.5-397b-a17b","input_price":0.5,"max_output_tokens":235929,"name":"Qwen: Qwen3.5 397B A17B","new_model":false,"open_weights":true,"output_price":3.6,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2026-02-15","score":1441.6936982868262,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3.5-397b-a17b-20260216","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Qwen3.5-397B-A17B","confidence_interval":[1438.6290883691815,1444.7583082044712],"vote_count":96206.0,"benchmark_stale":false,"listed_days_ago":253,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"qwen/qwen3.5-122b-a10b-20260224","catalog_created_at":"2026-02-25T21:09:49+00:00","catalog_status":"active","context_length":262144,"eligible":true,"extra_pricing":{},"id":"qwen/qwen3.5-122b-a10b","input_price":0.4,"max_output_tokens":65536,"name":"Qwen: Qwen3.5-122B-A10B","new_model":false,"open_weights":true,"output_price":3.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2026-02-23","score":1416.567975329745,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3.5-122b-a10b-20260224","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Qwen3.5-122B-A10B","confidence_interval":[1412.2825757814721,1420.853374878018],"vote_count":29716.0,"benchmark_stale":false,"listed_days_ago":245,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.27,"canonical_slug":"qwen/qwen3.5-27b-20260224","catalog_created_at":"2026-02-25T21:10:10+00:00","catalog_status":"active","context_length":262144,"eligible":true,"extra_pricing":{},"id":"qwen/qwen3.5-27b","input_price":0.27,"max_output_tokens":65536,"name":"Qwen: Qwen3.5-27B","new_model":false,"open_weights":true,"output_price":2.16,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2026-02-23","score":1408.724565227215,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3.5-27b-20260224","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Qwen3.5-27B","confidence_interval":[1404.433984361834,1413.015146092596],"vote_count":28661.0,"benchmark_stale":false,"listed_days_ago":245,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.05,"canonical_slug":"qwen/qwen3.5-35b-a3b-20260224","catalog_created_at":"2026-02-25T21:10:22+00:00","catalog_status":"active","context_length":262144,"eligible":true,"extra_pricing":{},"id":"qwen/qwen3.5-35b-a3b","input_price":0.14,"max_output_tokens":81920,"name":"Qwen: Qwen3.5-35B-A3B","new_model":false,"open_weights":true,"output_price":1.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2026-02-23","score":1393.9437750184882,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3.5-35b-a3b-20260224","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Qwen3.5-35B-A3B","confidence_interval":[1389.7126041478036,1398.1749458891727],"vote_count":30443.0,"benchmark_stale":false,"listed_days_ago":245,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"qwen/qwen3.5-flash-20260224","catalog_created_at":"2026-02-25T21:09:36+00:00","catalog_status":"active","context_length":1000000,"eligible":true,"extra_pricing":{},"id":"qwen/qwen3.5-flash-02-23","input_price":0.065,"max_output_tokens":65536,"name":"Qwen: Qwen3.5-Flash","new_model":false,"open_weights":false,"output_price":0.26,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2026-02-25","score":1395.9604829795883,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3.5-flash-20260224","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"date","mapping_note":"OpenRouter 路由 qwen/qwen3.5-flash-02-23 带版本日期，Arena 条目未标日期，按同一型号对应","source_model_id":"Qwen3.5-Flash","confidence_interval":[1392.296146188039,1399.6248197711377],"vote_count":61802.0,"benchmark_stale":false,"listed_days_ago":243,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"qwen/qwen3.6-plus-04-02","catalog_created_at":"2026-04-02T12:39:17+00:00","catalog_status":"active","context_length":1000000,"eligible":true,"extra_pricing":{},"id":"qwen/qwen3.6-plus","input_price":0.325,"max_output_tokens":65536,"name":"Qwen: Qwen3.6 Plus","new_model":false,"open_weights":false,"output_price":1.95,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2026-04-02","score":1443.1174163752908,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3.6-plus-04-02","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Qwen3.6-Plus","confidence_interval":[1438.9847573382924,1447.250075412289],"vote_count":48142.0,"benchmark_stale":false,"listed_days_ago":207,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.064,"canonical_slug":"qwen/qwen3.7-plus-20260602","catalog_created_at":"2026-06-03T13:03:03+00:00","catalog_status":"active","context_length":1000000,"eligible":true,"extra_pricing":{},"id":"qwen/qwen3.7-plus","input_price":0.32,"max_output_tokens":131072,"name":"Qwen: Qwen3.7 Plus","new_model":false,"open_weights":false,"output_price":1.28,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2026-06-02","score":1455.7042471746377,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3.7-plus-20260602","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":148.82,"ci_low":145.25,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-06-02"],"catalog_release_date":"2026-06-02","source_model_date":"2026-06-02","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"qwen/qwen3.7-plus-20260602","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":72,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":147.0,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-06-02","source_model_id":"Qwen3.7-Plus","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Qwen3.7-Plus","confidence_interval":[1451.3763931612818,1460.0321011879935],"vote_count":42058.0,"benchmark_stale":false,"listed_days_ago":146,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.09,"canonical_slug":"qwen/qwen3.8-27b-20260814","catalog_created_at":"2026-08-14T15:55:10+00:00","catalog_status":"active","context_length":1000000,"eligible":true,"extra_pricing":{},"id":"qwen/qwen3.8-27b","input_price":0.2,"max_output_tokens":235929,"name":"Qwen: Qwen3.8 27B","new_model":false,"open_weights":true,"output_price":2.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2026-08-14","score":1437.138461278535,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3.8-27b-20260814","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":153.72,"ci_low":149.75,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-08-14"],"catalog_release_date":"2026-08-14","source_model_date":"2026-08-14","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"qwen/qwen3.8-27b-20260814","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":54,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":151.69,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-08-14","source_model_id":"Qwen 3.8 27B","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Qwen3.8-27B","confidence_interval":[1432.2144339002552,1442.062488656815],"vote_count":26238.0,"benchmark_stale":false,"listed_days_ago":73,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.25,"canonical_slug":"qwen/qwen3.8-max-20260902","catalog_created_at":"2026-09-03T21:08:24+00:00","catalog_status":"active","context_length":1000000,"eligible":true,"extra_pricing":{},"id":"qwen/qwen3.8-max-0902","input_price":2.0,"max_output_tokens":131072,"name":"Qwen: Qwen3.8 Max (0902)","new_model":false,"open_weights":false,"output_price":6.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2026-09-02","score":1482.9118943499611,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3.8-max-20260902","version_policy":"independent_snapshot","evaluation_references":[{"benchmark":"Vals Index","benchmark_version":"2.2","checked_at":"2026-10-26T00:20:45.595508+00:00","date":"2026-10-24","effort":null,"identity_evidence":{"checked_at":"2026-10-26T00:20:45.595508+00:00","effort":null,"release_date":"2026-09-02","sha256":"ca570b6d36981e3c8862984fcde669077b08c7005a54fd0baf72c6d4736f2912","source_name":"Qwen 3.8 Max 0902","url":"https://www.vals.ai/models/alibaba_qwen3.8-max-0902"},"kind":"independent","model_url":"https://www.vals.ai/models/alibaba_qwen3.8-max-0902","model_version":"qwen/qwen3.8-max-20260902","note":"按经济领域权重汇总的任务成绩；原榜含部分其他模型使用备用模型的结果。推理档位为来源模型页默认值，单项测试可能不同，完整条件见原榜。","rank":11,"scope":"该版本综合榜全部已评测模型","score":59.05,"source_id":"vals","source_model_id":"Qwen 3.8 Max 0902","source_url":"https://www.vals.ai/benchmarks/vals_index","total":27,"unit":"%","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"date","mapping_note":"OpenRouter 路由 qwen/qwen3.8-max-0902 带版本日期，Arena 条目未标日期，按同一型号对应","source_model_id":"Qwen3.8-Max","confidence_interval":[1478.074341056792,1487.7494476431302],"vote_count":30500.0,"benchmark_stale":false,"listed_days_ago":54,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"stepfun/step-3.5-flash","catalog_created_at":"2026-01-29T23:12:17+00:00","catalog_status":"active","context_length":262144,"eligible":true,"extra_pricing":{},"id":"stepfun/step-3.5-flash","input_price":0.1,"max_output_tokens":65536,"name":"StepFun: Step 3.5 Flash","new_model":false,"open_weights":true,"output_price":0.3,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"stepfun","provider_label":"stepfun","release_date":"2026-01-29","score":1392.9703531883067,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"stepfun/step-3.5-flash","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Step-3.5-Flash","confidence_interval":[1389.418730266936,1396.5219761096773],"vote_count":60329.0,"benchmark_stale":false,"listed_days_ago":270,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.1,"canonical_slug":"stepfun/step-5-preview-20261008","catalog_created_at":"2026-10-08T12:34:09+00:00","catalog_status":"active","context_length":1048576,"eligible":true,"extra_pricing":{},"id":"stepfun/step-5-preview","input_price":1.0,"max_output_tokens":64000,"name":"StepFun: Step 5 Preview","new_model":false,"open_weights":true,"output_price":2.7,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"stepfun","provider_label":"stepfun","release_date":"2026-09-16","score":1447.682422338529,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"stepfun/step-5-preview-20261008","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Step 5 Preview","confidence_interval":[1441.0219686129528,1454.3428760641054],"vote_count":9792.0,"benchmark_stale":false,"listed_days_ago":40,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.020625,"canonical_slug":"tencent/hy3-20260706","catalog_created_at":"2026-07-06T13:20:48+00:00","catalog_status":"active","context_length":262144,"eligible":true,"extra_pricing":{},"id":"tencent/hy3","input_price":0.0825,"max_output_tokens":128000,"name":"Tencent: Hy3","new_model":false,"open_weights":true,"output_price":0.33,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"tencent","provider_label":"tencent","release_date":"2026-07-06","score":1453.7970900098853,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"tencent/hy3-20260706","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Hy3","confidence_interval":[1447.810100904093,1459.7840791156777],"vote_count":13220.0,"benchmark_stale":false,"listed_days_ago":112,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.06,"canonical_slug":"tencent/hy3-preview-20260421","catalog_created_at":"2026-04-22T17:15:50+00:00","catalog_status":"active","context_length":262144,"eligible":true,"extra_pricing":{},"id":"tencent/hy3-preview","input_price":0.18,"max_output_tokens":235929,"name":"Tencent: Hy3 preview","new_model":false,"open_weights":true,"output_price":0.6,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"tencent","provider_label":"tencent","release_date":"2026-04-20","score":1409.0082849192663,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"tencent/hy3-preview-20260421","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Hy3 preview","confidence_interval":[1401.599065072127,1416.4175047664057],"vote_count":6875.0,"benchmark_stale":false,"listed_days_ago":189,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.17,"canonical_slug":"thinkingmachines/inkling-20260715","catalog_created_at":"2026-07-17T22:05:56+00:00","catalog_status":"active","context_length":524288,"eligible":true,"extra_pricing":{},"id":"thinkingmachines/inkling","input_price":1.0,"max_output_tokens":471859,"name":"Thinking Machines: Inkling","new_model":false,"open_weights":true,"output_price":4.05,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"thinkingmachines","provider_label":"thinkingmachines","release_date":"2026-07-15","score":1441.5997781740728,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"thinkingmachines/inkling-20260715","version_policy":"independent_snapshot","evaluation_references":[{"benchmark":"Vals Index","benchmark_version":"2.2","checked_at":"2026-10-26T00:20:45.595508+00:00","date":"2026-10-24","effort":"0","identity_evidence":{"checked_at":"2026-10-26T00:20:45.595508+00:00","effort":"0","release_date":"2026-07-15","sha256":"e5869ecfaa80195bf0708f36d46938983202d53ba85d41ea648f3871369e80bb","source_name":"Inkling","url":"https://www.vals.ai/models/thinkingmachines_inkling"},"kind":"independent","model_url":"https://www.vals.ai/models/thinkingmachines_inkling","model_version":"thinkingmachines/inkling-20260715","note":"按经济领域权重汇总的任务成绩；原榜含部分其他模型使用备用模型的结果。推理档位为来源模型页默认值，单项测试可能不同，完整条件见原榜。","rank":26,"scope":"该版本综合榜全部已评测模型","score":28.96,"source_id":"vals","source_model_id":"Inkling","source_url":"https://www.vals.ai/benchmarks/vals_index","total":27,"unit":"%","update_mode":"automatic","stale":false},{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":150.21,"ci_low":145.96,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-07-15"],"catalog_release_date":"2026-07-15","source_model_date":"2026-07-15","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"thinkingmachines/inkling-20260715","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":68,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":148.41,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-07-15","source_model_id":"Inkling","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Inkling","confidence_interval":[1437.3618510362164,1445.837705311929],"vote_count":43705.0,"benchmark_stale":false,"listed_days_ago":103,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.1,"canonical_slug":"thinkingmachines/inkling-small-20260730","catalog_created_at":"2026-07-30T20:25:17+00:00","catalog_status":"active","context_length":524288,"eligible":true,"extra_pricing":{},"id":"thinkingmachines/inkling-small","input_price":0.45,"max_output_tokens":262144,"name":"Thinking Machines: Inkling Small","new_model":false,"open_weights":true,"output_price":1.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"thinkingmachines","provider_label":"thinkingmachines","release_date":"2026-07-30","score":1404.5495117878907,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"thinkingmachines/inkling-small-20260730","version_policy":"independent_snapshot","evaluation_references":[{"benchmark":"Vals Index","benchmark_version":"2.2","checked_at":"2026-10-26T00:20:45.595508+00:00","date":"2026-10-24","effort":"0","identity_evidence":{"checked_at":"2026-10-26T00:20:45.595508+00:00","effort":"0","release_date":"2026-07-30","sha256":"a7081c15822570a1171b2f182a36141b9b5df44e864fb86990defbac93f73ce1","source_name":"Inkling Small","url":"https://www.vals.ai/models/thinkingmachines_inkling-small"},"kind":"independent","model_url":"https://www.vals.ai/models/thinkingmachines_inkling-small","model_version":"thinkingmachines/inkling-small-20260730","note":"按经济领域权重汇总的任务成绩；原榜含部分其他模型使用备用模型的结果。推理档位为来源模型页默认值，单项测试可能不同，完整条件见原榜。","rank":27,"scope":"该版本综合榜全部已评测模型","score":27.59,"source_id":"vals","source_model_id":"Inkling Small","source_url":"https://www.vals.ai/benchmarks/vals_index","total":27,"unit":"%","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Inkling-Small","confidence_interval":[1399.9987338490546,1409.1002897267267],"vote_count":35359.0,"benchmark_stale":false,"listed_days_ago":88,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.2,"canonical_slug":"x-ai/grok-4.3-20260430","catalog_created_at":"2026-04-30T23:30:21+00:00","catalog_status":"active","context_length":1000000,"eligible":true,"extra_pricing":{},"id":"x-ai/grok-4.3","input_price":1.25,"max_output_tokens":900000,"name":"SpaceXAI: Grok 4.3","new_model":false,"open_weights":false,"output_price":2.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"x-ai","provider_label":"x-ai","release_date":"2026-04-17","score":1441.6651712905812,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"x-ai/grok-4.3-20260430","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Grok 4.3","confidence_interval":[1438.0403065450205,1445.2900360361418],"vote_count":71348.0,"benchmark_stale":false,"listed_days_ago":192,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.3,"canonical_slug":"x-ai/grok-4.5-20260708","catalog_created_at":"2026-07-08T15:05:54+00:00","catalog_status":"active","context_length":500000,"eligible":true,"extra_pricing":{},"id":"x-ai/grok-4.5","input_price":2.0,"max_output_tokens":450000,"name":"SpaceXAI: Grok 4.5","new_model":false,"open_weights":false,"output_price":6.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"x-ai","provider_label":"x-ai","release_date":"2026-07-08","score":1465.7682811634877,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"x-ai/grok-4.5-20260708","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":156.18,"ci_low":152.36,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-07-08"],"catalog_release_date":"2026-07-08","source_model_date":"2026-07-08","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"x-ai/grok-4.5-20260708","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":43,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":154.07,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-07-08","source_model_id":"Grok 4.5","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"Grok 4.5","confidence_interval":[1461.4350086638426,1470.1015536631328],"vote_count":39880.0,"benchmark_stale":false,"listed_days_ago":110,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.5,"canonical_slug":"x-ai/grok-4.6-20260810","catalog_created_at":"2026-08-12T15:35:57+00:00","catalog_status":"active","context_length":500000,"eligible":true,"extra_pricing":{},"id":"x-ai/grok-4.6","input_price":2.0,"max_output_tokens":450000,"name":"SpaceXAI: Grok 4.6","new_model":false,"open_weights":false,"output_price":6.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"x-ai","provider_label":"x-ai","release_date":"2026-08-12","score":1456.3622941261822,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"x-ai/grok-4.6-20260810","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"high","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"high","mapping_type":"effort","mapping_note":"Arena 只收录了 high 档位；调用时需设置同一推理档位","source_model_id":"Grok 4.6 (High)","confidence_interval":[1451.6913530079987,1461.0332352443656],"vote_count":32189.0,"benchmark_stale":false,"listed_days_ago":75,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.5,"canonical_slug":"x-ai/grok-4.7-20260916","catalog_created_at":"2026-09-21T16:19:01+00:00","catalog_status":"active","context_length":500000,"eligible":true,"extra_pricing":{},"id":"x-ai/grok-4.7","input_price":2.0,"max_output_tokens":450000,"name":"SpaceXAI: Grok 4.7","new_model":false,"open_weights":false,"output_price":6.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"x-ai","provider_label":"x-ai","release_date":"2026-09-21","score":1441.0413192045798,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"x-ai/grok-4.7-20260916","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"xhigh","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"xhigh","mapping_type":"effort","mapping_note":"Arena 只收录了 xhigh 档位；调用时需设置同一推理档位","source_model_id":"Grok 4.7 (xHigh)","confidence_interval":[1435.1188645145749,1446.963773894585],"vote_count":14693.0,"benchmark_stale":false,"listed_days_ago":35,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.00255,"canonical_slug":"xiaomi/mimo-v2.5-20260422","catalog_created_at":"2026-04-22T16:11:09+00:00","catalog_status":"active","context_length":1050000,"eligible":true,"extra_pricing":{},"id":"xiaomi/mimo-v2.5","input_price":0.119,"max_output_tokens":945000,"name":"Xiaomi: MiMo-V2.5","new_model":false,"open_weights":true,"output_price":0.238,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"xiaomi","provider_label":"xiaomi","release_date":"2026-04-22","score":1433.9706183756462,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"xiaomi/mimo-v2.5-20260422","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"MiMo-V2.5","confidence_interval":[1429.749146476325,1438.1920902749675],"vote_count":47364.0,"benchmark_stale":false,"listed_days_ago":187,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.0028,"canonical_slug":"xiaomi/mimo-v2.5-pro-20260422","catalog_created_at":"2026-04-22T16:11:13+00:00","catalog_status":"active","context_length":1050000,"eligible":true,"extra_pricing":{},"id":"xiaomi/mimo-v2.5-pro","input_price":0.3045,"max_output_tokens":945000,"name":"Xiaomi: MiMo-V2.5-Pro","new_model":false,"open_weights":true,"output_price":0.609,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"xiaomi","provider_label":"xiaomi","release_date":"2026-04-22","score":1466.4622105406606,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"xiaomi/mimo-v2.5-pro-20260422","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"MiMo-V2.5-Pro","confidence_interval":[1462.9552553277838,1469.9691657535377],"vote_count":77883.0,"benchmark_stale":false,"listed_days_ago":187,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.0028,"canonical_slug":"xiaomi/mimo-v2.6-flash-20260921","catalog_created_at":"2026-09-21T20:07:44+00:00","catalog_status":"active","context_length":1050000,"eligible":true,"extra_pricing":{},"id":"xiaomi/mimo-v2.6-flash","input_price":0.14,"max_output_tokens":131072,"name":"Xiaomi: MiMo-V2.6-Flash","new_model":false,"open_weights":true,"output_price":0.28,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"xiaomi","provider_label":"xiaomi","release_date":"2026-09-22","score":1448.1035716010952,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"xiaomi/mimo-v2.6-flash-20260921","version_policy":"independent_snapshot","evaluation_references":[{"benchmark":"Vals Index","benchmark_version":"2.2","checked_at":"2026-10-26T00:20:45.595508+00:00","date":"2026-10-24","effort":null,"identity_evidence":{"checked_at":"2026-10-26T00:20:45.595508+00:00","effort":null,"release_date":"2026-09-21","sha256":"5df5d91825ac3d46c229ad401d338e9a96d0f2d47687437858dcbfbf2e8184c5","source_name":"MiMo V2.6 Flash","url":"https://www.vals.ai/models/xiaomi_mimo-v2.6-flash"},"kind":"independent","model_url":"https://www.vals.ai/models/xiaomi_mimo-v2.6-flash","model_version":"xiaomi/mimo-v2.6-flash-20260921","note":"按经济领域权重汇总的任务成绩；原榜含部分其他模型使用备用模型的结果。推理档位为来源模型页默认值，单项测试可能不同，完整条件见原榜。","rank":16,"scope":"该版本综合榜全部已评测模型","score":54.83,"source_id":"vals","source_model_id":"MiMo V2.6 Flash","source_url":"https://www.vals.ai/benchmarks/vals_index","total":27,"unit":"%","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"MiMo-V2.6-Flash","confidence_interval":[1442.0921150553786,1454.115028146812],"vote_count":13730.0,"benchmark_stale":false,"listed_days_ago":34,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.0036,"canonical_slug":"xiaomi/mimo-v2.6-pro-20260921","catalog_created_at":"2026-09-21T20:07:39+00:00","catalog_status":"active","context_length":1050000,"eligible":true,"extra_pricing":{},"id":"xiaomi/mimo-v2.6-pro","input_price":0.435,"max_output_tokens":131072,"name":"Xiaomi: MiMo-V2.6-Pro","new_model":false,"open_weights":true,"output_price":0.87,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"xiaomi","provider_label":"xiaomi","release_date":"2026-09-22","score":1479.5470990311535,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"xiaomi/mimo-v2.6-pro-20260921","version_policy":"independent_snapshot","evaluation_references":[{"benchmark":"Vals Index","benchmark_version":"2.2","checked_at":"2026-10-26T00:20:45.595508+00:00","date":"2026-10-24","effort":null,"identity_evidence":{"checked_at":"2026-10-26T00:20:45.595508+00:00","effort":null,"release_date":"2026-09-21","sha256":"170db53daa26de841d0b51d5b72f4683a229f1177fec42f2069ee0e07f57acad","source_name":"MiMo V2.6 Pro","url":"https://www.vals.ai/models/xiaomi_mimo-v2.6-pro"},"kind":"independent","model_url":"https://www.vals.ai/models/xiaomi_mimo-v2.6-pro","model_version":"xiaomi/mimo-v2.6-pro-20260921","note":"按经济领域权重汇总的任务成绩；原榜含部分其他模型使用备用模型的结果。推理档位为来源模型页默认值，单项测试可能不同，完整条件见原榜。","rank":14,"scope":"该版本综合榜全部已评测模型","score":57.81,"source_id":"vals","source_model_id":"MiMo V2.6 Pro","source_url":"https://www.vals.ai/benchmarks/vals_index","total":27,"unit":"%","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"MiMo-V2.6-Pro","confidence_interval":[1470.1669744871922,1488.9272235751148],"vote_count":4067.0,"benchmark_stale":false,"listed_days_ago":34,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.11,"canonical_slug":"z-ai/glm-4.5","catalog_created_at":"2025-07-25T19:22:27+00:00","catalog_status":"active","context_length":131072,"eligible":true,"extra_pricing":{},"id":"z-ai/glm-4.5","input_price":0.6,"max_output_tokens":98304,"name":"Z.ai: GLM 4.5","new_model":false,"open_weights":true,"output_price":2.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"z-ai","provider_label":"z-ai","release_date":"2025-07-28","score":1410.924045244777,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"z-ai/glm-4.5","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"GLM-4.5","confidence_interval":[1406.001277007957,1415.8468134815973],"vote_count":23791.0,"benchmark_stale":false,"listed_days_ago":455,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.025,"canonical_slug":"z-ai/glm-4.5-air","catalog_created_at":"2025-07-25T19:20:58+00:00","catalog_status":"active","context_length":131072,"eligible":true,"extra_pricing":{},"id":"z-ai/glm-4.5-air","input_price":0.13,"max_output_tokens":98304,"name":"Z.ai: GLM 4.5 Air","new_model":false,"open_weights":true,"output_price":0.85,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"z-ai","provider_label":"z-ai","release_date":"2025-07-28","score":1373.0059375462004,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"z-ai/glm-4.5-air","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"GLM-4.5-Air","confidence_interval":[1368.7521857881572,1377.2596893042435],"vote_count":30646.0,"benchmark_stale":false,"listed_days_ago":455,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.11,"canonical_slug":"z-ai/glm-4.5v","catalog_created_at":"2025-08-11T14:24:48+00:00","catalog_status":"active","context_length":65536,"eligible":true,"extra_pricing":{},"id":"z-ai/glm-4.5v","input_price":0.6,"max_output_tokens":16384,"name":"Z.ai: GLM 4.5V","new_model":false,"open_weights":true,"output_price":1.8,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"z-ai","provider_label":"z-ai","release_date":"2025-08-11","score":1352.1670977462725,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"z-ai/glm-4.5v","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"GLM-4.5V","confidence_interval":[1343.73085714207,1360.603338350475],"vote_count":4821.0,"benchmark_stale":false,"listed_days_ago":441,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.11,"canonical_slug":"z-ai/glm-4.6","catalog_created_at":"2025-09-30T12:32:56+00:00","catalog_status":"active","context_length":204800,"eligible":true,"extra_pricing":{},"id":"z-ai/glm-4.6","input_price":0.55,"max_output_tokens":131072,"name":"Z.ai: GLM 4.6","new_model":false,"open_weights":true,"output_price":2.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"z-ai","provider_label":"z-ai","release_date":"2025-09-30","score":1424.3380567750482,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"z-ai/glm-4.6","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"GLM-4.6","confidence_interval":[1420.4375432086497,1428.2385703414468],"vote_count":35818.0,"benchmark_stale":false,"listed_days_ago":391,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.05,"canonical_slug":"z-ai/glm-4.6-20251208","catalog_created_at":"2025-12-08T15:24:22+00:00","catalog_status":"active","context_length":131072,"eligible":true,"extra_pricing":{},"id":"z-ai/glm-4.6v","input_price":0.3,"max_output_tokens":32768,"name":"Z.ai: GLM 4.6V","new_model":false,"open_weights":true,"output_price":0.9,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"z-ai","provider_label":"z-ai","release_date":"2025-12-08","score":1378.2282643648057,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"z-ai/glm-4.6-20251208","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"GLM-4.6V","confidence_interval":[1367.0443601921502,1389.4121685374612],"vote_count":2831.0,"benchmark_stale":false,"listed_days_ago":322,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.099,"canonical_slug":"z-ai/glm-4.7-20251222","catalog_created_at":"2025-12-22T04:33:34+00:00","catalog_status":"active","context_length":204800,"eligible":true,"extra_pricing":{},"id":"z-ai/glm-4.7","input_price":0.54,"max_output_tokens":131072,"name":"Z.ai: GLM 4.7","new_model":false,"open_weights":true,"output_price":1.98,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"z-ai","provider_label":"z-ai","release_date":"2025-12-22","score":1441.3054568978723,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"z-ai/glm-4.7-20251222","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":145.85,"ci_low":141.38,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-12-22"],"catalog_release_date":"2025-12-22","source_model_date":"2025-12-22","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"z-ai/glm-4.7-20251222","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":100,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":143.59,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-12-22","source_model_id":"GLM-4.7","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"GLM-4.7","confidence_interval":[1435.2481408117874,1447.3627729839573],"vote_count":12245.0,"benchmark_stale":false,"listed_days_ago":308,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":null,"canonical_slug":"z-ai/glm-4.7-flash-20260119","catalog_created_at":"2026-01-19T14:45:13+00:00","catalog_status":"active","context_length":200000,"eligible":true,"extra_pricing":{},"id":"z-ai/glm-4.7-flash","input_price":0.0605,"max_output_tokens":117964,"name":"Z.ai: GLM 4.7 Flash","new_model":false,"open_weights":true,"output_price":0.4,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"z-ai","provider_label":"z-ai","release_date":"2026-01-19","score":1364.7191823451406,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"z-ai/glm-4.7-flash-20260119","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"GLM-4.7-Flash","confidence_interval":[1359.0061816631817,1370.4321830270997],"vote_count":12013.0,"benchmark_stale":false,"listed_days_ago":280,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.2,"canonical_slug":"z-ai/glm-5-20260211","catalog_created_at":"2026-02-11T16:59:42+00:00","catalog_status":"active","context_length":204800,"eligible":true,"extra_pricing":{},"id":"z-ai/glm-5","input_price":1.0,"max_output_tokens":131072,"name":"Z.ai: GLM 5","new_model":false,"open_weights":true,"output_price":3.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"z-ai","provider_label":"z-ai","release_date":"2026-02-12","score":1457.4128976153304,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"z-ai/glm-5-20260211","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"GLM-5","confidence_interval":[1453.170016615211,1461.65577861545],"vote_count":29267.0,"benchmark_stale":false,"listed_days_ago":256,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.26,"canonical_slug":"z-ai/glm-5.1-20260406","catalog_created_at":"2026-04-07T16:07:05+00:00","catalog_status":"active","context_length":204800,"eligible":true,"extra_pricing":{},"id":"z-ai/glm-5.1","input_price":1.4,"max_output_tokens":131072,"name":"Z.ai: GLM 5.1","new_model":false,"open_weights":true,"output_price":4.4,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"z-ai","provider_label":"z-ai","release_date":"2026-04-07","score":1465.3593126717285,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"z-ai/glm-5.1-20260406","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"GLM-5.1","confidence_interval":[1461.812592560784,1468.906032782673],"vote_count":65402.0,"benchmark_stale":false,"listed_days_ago":202,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.033,"canonical_slug":"z-ai/glm-5.2-20260616","catalog_created_at":"2026-06-16T17:45:30+00:00","catalog_status":"active","context_length":1048576,"eligible":true,"extra_pricing":{},"id":"z-ai/glm-5.2","input_price":0.0341,"max_output_tokens":943718,"name":"Z.ai: GLM 5.2","new_model":false,"open_weights":true,"output_price":7.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"z-ai","provider_label":"z-ai","release_date":"2026-06-13","score":1475.4156684253448,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"z-ai/glm-5.2-20260616","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"max","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"max","mapping_type":"effort","mapping_note":"Arena 只收录了 max 档位；调用时需设置同一推理档位","source_model_id":"GLM-5.2 (Max)","confidence_interval":[1471.3017166265654,1479.5296202241243],"vote_count":52283.0,"benchmark_stale":false,"listed_days_ago":135,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.078,"canonical_slug":"z-ai/glm-5.3-20260816","catalog_created_at":"2026-08-18T20:57:35+00:00","catalog_status":"active","context_length":1048576,"eligible":true,"extra_pricing":{},"id":"z-ai/glm-5.3","input_price":0.079,"max_output_tokens":943718,"name":"Z.ai: GLM 5.3","new_model":false,"open_weights":true,"output_price":1.4,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"z-ai","provider_label":"z-ai","release_date":"2026-08-14","score":1480.7355272301697,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"z-ai/glm-5.3-20260816","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"max","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":"max","mapping_type":"effort","mapping_note":"Arena 只收录了 max 档位；调用时需设置同一推理档位","source_model_id":"GLM-5.3 (Max)","confidence_interval":[1475.7041682312126,1485.766886229127],"vote_count":25144.0,"benchmark_stale":false,"listed_days_ago":73,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.03,"canonical_slug":"z-ai/glm-5.3-flash-20260826","catalog_created_at":"2026-08-26T13:59:01+00:00","catalog_status":"active","context_length":1048576,"eligible":true,"extra_pricing":{},"id":"z-ai/glm-5.3-flash","input_price":0.15,"max_output_tokens":943717,"name":"Z.ai: GLM 5.3 Flash","new_model":false,"open_weights":true,"output_price":0.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"z-ai","provider_label":"z-ai","release_date":"2026-08-26","score":1475.957904524776,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"z-ai/glm-5.3-flash-20260826","version_policy":"independent_snapshot","evaluation_references":[{"benchmark":"Vals Index","benchmark_version":"2.2","checked_at":"2026-10-26T00:20:45.595508+00:00","date":"2026-10-24","effort":"max","identity_evidence":{"checked_at":"2026-10-26T00:20:45.595508+00:00","effort":"max","release_date":"2026-08-26","sha256":"fd790cd563635a77a0c10777b08f676910ce7e9b316b07e3f6ec975990661383","source_name":"GLM 5.3 Flash","url":"https://www.vals.ai/models/zai_glm-5.3-flash"},"kind":"independent","model_url":"https://www.vals.ai/models/zai_glm-5.3-flash","model_version":"z-ai/glm-5.3-flash-20260826","note":"按经济领域权重汇总的任务成绩；原榜含部分其他模型使用备用模型的结果。推理档位为来源模型页默认值，单项测试可能不同，完整条件见原榜。","rank":22,"scope":"该版本综合榜全部已评测模型","score":50.86,"source_id":"vals","source_model_id":"GLM 5.3 Flash","source_url":"https://www.vals.ai/benchmarks/vals_index","total":27,"unit":"%","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"GLM-5.3-Flash","confidence_interval":[1471.2361238601075,1480.6796851894446],"vote_count":38359.0,"benchmark_stale":false,"listed_days_ago":61,"new_model_basis":"release_date","variant":false},{"benchmark":"arena/text_style_control/overall","benchmark_date":"2026-10-22","benchmark_source":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset/viewer/text_style_control/latest","benchmark_status":"source_snapshot","benchmark_version":"2026-10-22","cache_read_price":0.24,"canonical_slug":"z-ai/glm-5v-turbo-20260401","catalog_created_at":"2026-04-01T16:37:38+00:00","catalog_status":"active","context_length":202752,"eligible":true,"extra_pricing":{},"id":"z-ai/glm-5v-turbo","input_price":1.2,"max_output_tokens":131072,"name":"Z.ai: GLM 5V Turbo","new_model":false,"open_weights":false,"output_price":4.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"z-ai","provider_label":"z-ai","release_date":"2026-04-01","score":1433.9223996040687,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"z-ai/glm-5v-turbo-20260401","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"},{"id":"overall","label":"综合","date":"2026-10-22"},{"id":"chinese","label":"中文","date":"2026-10-22"},{"id":"coding","label":"写代码","date":"2026-10-22"},{"id":"creative_writing","label":"写作","date":"2026-10-22"},{"id":"hard_prompts","label":"复杂难题","date":"2026-10-22"},{"id":"instruction_following","label":"按格式办事","date":"2026-10-22"},{"id":"longer_query","label":"长文本","date":"2026-10-22"},{"id":"math","label":"数学","date":"2026-10-22"},{"id":"multi_turn","label":"多轮对话","date":"2026-10-22"}],"benchmark_scale_max":3000.0,"benchmark_configuration":"default","score_unit":"Arena","score_semantics":"Arena Bradley-Terry 评分（Elo 尺度）。分差可换算为人类盲测的预期胜率，分数比例没有意义。","benchmark_category":"overall","effort":null,"mapping_type":"exact","mapping_note":"Arena 名称与 OpenRouter 路由 ID 一致","source_model_id":"GLM-5V-Turbo","confidence_interval":[1427.4901328244273,1440.3546663837103],"vote_count":10477.0,"benchmark_stale":false,"listed_days_ago":208,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.003,"canonical_slug":"deepseek/deepseek-v4.1-flash-20261017","catalog_created_at":"2026-10-17T08:27:04+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"deepseek/deepseek-v4.1-flash-1017","input_price":0.15,"max_output_tokens":393216,"name":"DeepSeek: DeepSeek V4.1 Flash 1017","new_model":true,"open_weights":true,"output_price":0.6,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"deepseek","provider_label":"deepseek","release_date":"2026-10-17","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"official_release":{"announcement_url":"https://x.com/deepseek_ai/status/2111375640717033777","api_alias":"deepseek-flash","id":"deepseek/deepseek-v4.1-flash-1017","model_card_url":"https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash-1017","official_scores":[{"benchmark":"HLE (w/o tools)","date":"2026-10-17","kind":"vendor_reported","model_version":"DeepSeek-V4.1-Flash-1017","reported_value":"42.6 (44.0*)","score":42.6,"source_url":"https://api-docs.deepseek.com/updates/#deepseek-v41-flash-1017-update"},{"benchmark":"Terminal-Bench 4.0","date":"2026-10-17","kind":"vendor_reported","model_version":"DeepSeek-V4.1-Flash-1017","reported_value":"42.3","score":42.3,"source_url":"https://api-docs.deepseek.com/updates/#deepseek-v41-flash-1017-update"},{"benchmark":"DeepSWE v1.1","date":"2026-10-17","kind":"vendor_reported","model_version":"DeepSeek-V4.1-Flash-1017","reported_value":"77.0","score":77.0,"source_url":"https://api-docs.deepseek.com/updates/#deepseek-v41-flash-1017-update"},{"benchmark":"ProgramBench","date":"2026-10-17","kind":"vendor_reported","model_version":"DeepSeek-V4.1-Flash-1017","reported_value":"29.1","score":29.1,"source_url":"https://api-docs.deepseek.com/updates/#deepseek-v41-flash-1017-update"},{"benchmark":"NL2Repo-Bench","date":"2026-10-17","kind":"vendor_reported","model_version":"DeepSeek-V4.1-Flash-1017","reported_value":"68.2","score":68.2,"source_url":"https://api-docs.deepseek.com/updates/#deepseek-v41-flash-1017-update"},{"benchmark":"Automation-Bench v1.0.6","date":"2026-10-17","kind":"vendor_reported","model_version":"DeepSeek-V4.1-Flash-1017","reported_value":"60.5","score":60.5,"source_url":"https://api-docs.deepseek.com/updates/#deepseek-v41-flash-1017-update"},{"benchmark":"Agents' Last Exam","date":"2026-10-17","kind":"vendor_reported","model_version":"DeepSeek-V4.1-Flash-1017","reported_value":"35.4","score":35.4,"source_url":"https://api-docs.deepseek.com/updates/#deepseek-v41-flash-1017-update"},{"benchmark":"Apex Bench v1.1","date":"2026-10-17","kind":"vendor_reported","model_version":"DeepSeek-V4.1-Flash-1017","reported_value":"72.1","score":72.1,"source_url":"https://api-docs.deepseek.com/updates/#deepseek-v41-flash-1017-update"},{"benchmark":"Chartography (w/tools)","date":"2026-10-17","kind":"vendor_reported","model_version":"DeepSeek-V4.1-Flash-1017","reported_value":"83.0","score":83.0,"source_url":"https://api-docs.deepseek.com/updates/#deepseek-v41-flash-1017-update"},{"benchmark":"BabyVision (w/tools)","date":"2026-10-17","kind":"vendor_reported","model_version":"DeepSeek-V4.1-Flash-1017","reported_value":"92.5","score":92.5,"source_url":"https://api-docs.deepseek.com/updates/#deepseek-v41-flash-1017-update"},{"benchmark":"ZeroBench-main (w/tools)","date":"2026-10-17","kind":"vendor_reported","model_version":"DeepSeek-V4.1-Flash-1017","reported_value":"51.0","score":51.0,"source_url":"https://api-docs.deepseek.com/updates/#deepseek-v41-flash-1017-update"}],"release_date":"2026-10-17","source_url":"https://api-docs.deepseek.com/updates/#deepseek-v41-flash-1017-update","version":"DeepSeek-V4.1-Flash-1017","version_source_url":"https://api-docs.deepseek.com/quick_start/pricing/"},"official_source_stale":false,"model_version":"deepseek/deepseek-v4.1-flash-20261017","version_policy":"independent_snapshot","evaluation_references":[{"benchmark":"Vals Index","benchmark_version":"2.2","checked_at":"2026-10-26T00:20:45.595508+00:00","date":"2026-10-24","effort":"max","identity_evidence":{"checked_at":"2026-10-26T00:20:45.595508+00:00","effort":"max","release_date":"2026-10-17","sha256":"597c75ced2c458e0b35f062bc95b31b7b6ea51f122479edb896f657d6c318b50","source_name":"DeepSeek V4.1 Flash 1017","url":"https://www.vals.ai/models/deepseek_deepseek-v4.1-flash-1017"},"kind":"independent","model_url":"https://www.vals.ai/models/deepseek_deepseek-v4.1-flash-1017","model_version":"deepseek/deepseek-v4.1-flash-20261017","note":"按经济领域权重汇总的任务成绩；原榜含部分其他模型使用备用模型的结果。推理档位为来源模型页默认值，单项测试可能不同，完整条件见原榜。","rank":9,"scope":"该版本综合榜全部已评测模型","score":61.26,"source_id":"vals","source_model_id":"DeepSeek V4.1 Flash 1017","source_url":"https://www.vals.ai/benchmarks/vals_index","total":27,"unit":"%","update_mode":"automatic","stale":false}],"evaluation_history":[{"model_version":"deepseek/deepseek-v4.1-flash-20261017","source_model_id":"DeepSeek V4.1 Flash 1017 (Max)","benchmark":"Artificial Analysis Intelligence Index","benchmark_version":"4.3.2","score":45,"unit":"分","rank":2,"total":121,"scope":"模型页所列同类开源权重模型（非全模型总榜）","source_url":"https://artificialanalysis.ai/models/deepseek-v4-1-flash-1017","kind":"independent","date":"2026-10-23","checked_at":"2026-10-23T05:22:07+00:00","source_id":"aa_history","update_mode":"historical_snapshot","note":"2026-10-23 留存的历史快照，max 推理档，不代表当前成绩或名次。当前综合参考由自动更新的 Vals、Epoch 提供；AA 公开再分发需单独授权，本快照不设人工复核任务。"}],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"}],"effort":null,"mapping_type":null,"listed_days_ago":9,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.06,"canonical_slug":"elorian/glimpse-20261016","catalog_created_at":"2026-10-17T03:10:29+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"elorian/glimpse","input_price":0.6,"max_output_tokens":256000,"name":"Elorian: Glimpse","new_model":true,"open_weights":false,"output_price":3.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"elorian","provider_label":"elorian","release_date":"2026-10-17","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"elorian/glimpse-20261016","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":9,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"google/gemini-nano-banana-2.1-20261006","catalog_created_at":"2026-10-06T15:33:45+00:00","catalog_status":"active","context_length":65536,"eligible":false,"extra_pricing":{},"id":"google/gemini-nano-banana-2.1","input_price":1.5,"max_output_tokens":58982,"name":"Google: Nano Banana 2.1","new_model":true,"open_weights":false,"output_price":7.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2026-10-06","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-nano-banana-2.1-20261006","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":20,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.01,"canonical_slug":"inception/mercury-3-20261019","catalog_created_at":"2026-10-16T13:45:37+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"inception/mercury-3","input_price":0.1,"max_output_tokens":131072,"name":"Inception: Mercury 3","new_model":true,"open_weights":false,"output_price":0.35,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"inception","provider_label":"inception","release_date":"2026-10-16","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"inception/mercury-3-20261019","version_policy":"independent_snapshot","evaluation_references":[{"benchmark":"Vals Index","benchmark_version":"2.2","checked_at":"2026-10-26T00:20:45.595508+00:00","date":"2026-10-24","effort":"high","identity_evidence":{"checked_at":"2026-10-26T00:20:45.595508+00:00","effort":"high","release_date":"2026-10-19","sha256":"07abb9558e0d1425ed2aebeb57a171cca80845893e0b02661e2282c75559f464","source_name":"Mercury 3","url":"https://www.vals.ai/models/inception_mercury-3"},"kind":"independent","model_url":"https://www.vals.ai/models/inception_mercury-3","model_version":"inception/mercury-3-20261019","note":"按经济领域权重汇总的任务成绩；原榜含部分其他模型使用备用模型的结果。推理档位为来源模型页默认值，单项测试可能不同，完整条件见原榜。","rank":24,"scope":"该版本综合榜全部已评测模型","score":43.66,"source_id":"vals","source_model_id":"Mercury 3","source_url":"https://www.vals.ai/benchmarks/vals_index","total":27,"unit":"%","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":10,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.015,"canonical_slug":"inclusionai/ling-3.1-flash-20261002","catalog_created_at":"2026-10-02T14:07:04+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"inclusionai/ling-3.1-flash","input_price":0.075,"max_output_tokens":235929,"name":"inclusionAI: Ling 3.1 Flash","new_model":true,"open_weights":true,"output_price":0.225,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"inclusionai","provider_label":"inclusionai","release_date":"2026-09-29","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"inclusionai/ling-3.1-flash-20261002","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":27,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.025,"canonical_slug":"liquid/lfm-2.5-24b-a2b-20261012","catalog_created_at":"2026-10-13T17:20:25+00:00","catalog_status":"active","context_length":128000,"eligible":false,"extra_pricing":{},"id":"liquid/lfm-2.5-24b-a2b","input_price":0.05,"max_output_tokens":8192,"name":"LiquidAI: LFM2.5-24B-A2B","new_model":true,"open_weights":true,"output_price":0.7,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"liquid","provider_label":"liquid","release_date":"2026-10-13","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"liquid/lfm-2.5-24b-a2b-20261012","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":13,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.15,"canonical_slug":"meta/muse-spark-open-20261015","catalog_created_at":"2026-10-15T17:57:49+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"meta/muse-spark-open","input_price":1.25,"max_output_tokens":943718,"name":"Meta: Muse Spark Open","new_model":true,"open_weights":true,"output_price":4.25,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"meta","provider_label":"meta","release_date":"2026-10-15","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"meta/muse-spark-open-20261015","version_policy":"independent_snapshot","evaluation_references":[{"benchmark":"Vals Index","benchmark_version":"2.2","checked_at":"2026-10-26T00:20:45.595508+00:00","date":"2026-10-24","effort":"xhigh","identity_evidence":{"checked_at":"2026-10-26T00:20:45.595508+00:00","effort":"xhigh","release_date":"2026-10-15","sha256":"b5052f2cee6821b1a117c45d001dee78ae79a37cfa0b2abf2a71963d61b719a7","source_name":"Muse Spark Open","url":"https://www.vals.ai/models/meta_muse-spark-open"},"kind":"independent","model_url":"https://www.vals.ai/models/meta_muse-spark-open","model_version":"meta/muse-spark-open-20261015","note":"按经济领域权重汇总的任务成绩；原榜含部分其他模型使用备用模型的结果。推理档位为来源模型页默认值，单项测试可能不同，完整条件见原榜。","rank":23,"scope":"该版本综合榜全部已评测模型","score":48.57,"source_id":"vals","source_model_id":"Muse Spark Open","source_url":"https://www.vals.ai/benchmarks/vals_index","total":27,"unit":"%","update_mode":"automatic","stale":false},{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":157.97,"ci_low":152.66,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-10-15"],"catalog_release_date":"2026-10-15","source_model_date":"2026-10-15","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"meta/muse-spark-open-20261015","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":36,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":155.11,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-10-15","source_model_id":"Muse Spark Open","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"}],"effort":null,"mapping_type":null,"listed_days_ago":11,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.015,"canonical_slug":"minimax/minimax-m3.1-flash-20261019","catalog_created_at":"2026-10-19T03:12:49+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"minimax/minimax-m3.1-flash","input_price":0.15,"max_output_tokens":524288,"name":"MiniMax: MiniMax M3.1 Flash","new_model":true,"open_weights":false,"output_price":0.6,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"minimax","provider_label":"minimax","release_date":"2026-10-19","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"minimax/minimax-m3.1-flash-20261019","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"}],"effort":null,"mapping_type":null,"listed_days_ago":7,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.14,"canonical_slug":"mistralai/mistral-large-4-0-20261006","catalog_created_at":"2026-10-06T13:42:02+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"mistralai/mistral-large-4-0","input_price":1.36,"max_output_tokens":262144,"name":"Mistral: Mistral Large 4","new_model":true,"open_weights":false,"output_price":4.18,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"mistralai","provider_label":"mistralai","release_date":"2026-10-06","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"mistralai/mistral-large-4-0-20261006","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":152.74,"ci_low":146.94,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-10-06"],"catalog_release_date":"2026-10-06","source_model_date":"2026-10-06","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"mistralai/mistral-large-4-0-20261006","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":60,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":150.03,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-10-06","source_model_id":"Mistral Large 4","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":20,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"moondream/moondream3-preview-20261013","catalog_created_at":"2026-10-15T22:13:05+00:00","catalog_status":"active","context_length":32000,"eligible":false,"extra_pricing":{},"id":"moondream/moondream3-preview","input_price":0.3,"max_output_tokens":4096,"name":"Moondream: Moondream 3 Preview","new_model":true,"open_weights":true,"output_price":2.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"moondream","provider_label":"moondream","release_date":"2026-10-15","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"moondream/moondream3-preview-20261013","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":11,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"moondream/moondream3.1-9b-a2b-20261013","catalog_created_at":"2026-10-15T22:12:53+00:00","catalog_status":"active","context_length":32000,"eligible":false,"extra_pricing":{},"id":"moondream/moondream3.1-9b-a2b","input_price":0.3,"max_output_tokens":4096,"name":"Moondream: Moondream 3.1","new_model":true,"open_weights":true,"output_price":2.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"moondream","provider_label":"moondream","release_date":"2026-10-15","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"moondream/moondream3.1-9b-a2b-20261013","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":11,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.2,"canonical_slug":"moonshotai/kimi-k3.1-pro-20261023","catalog_created_at":"2026-10-23T17:56:28+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"moonshotai/kimi-k3.1-pro","input_price":2.0,"max_output_tokens":943718,"name":"MoonshotAI: Kimi K3.1 Pro","new_model":true,"open_weights":false,"output_price":10.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"moonshotai","provider_label":"moonshotai","release_date":"2026-10-23","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"moonshotai/kimi-k3.1-pro-20261023","version_policy":"independent_snapshot","evaluation_references":[{"benchmark":"Vals Index","benchmark_version":"2.2","checked_at":"2026-10-26T00:20:45.595508+00:00","date":"2026-10-24","effort":"max","identity_evidence":{"checked_at":"2026-10-26T00:20:45.595508+00:00","effort":"max","release_date":"2026-10-23","sha256":"28048bee6bfa60e8257402143a76ad25060458aae67d24b74c073d5df531977b","source_name":"Kimi K3.1 Pro","url":"https://www.vals.ai/models/kimi_kimi-k3.1-pro"},"kind":"independent","model_url":"https://www.vals.ai/models/kimi_kimi-k3.1-pro","model_version":"moonshotai/kimi-k3.1-pro-20261023","note":"按经济领域权重汇总的任务成绩；原榜含部分其他模型使用备用模型的结果。推理档位为来源模型页默认值，单项测试可能不同，完整条件见原榜。","rank":8,"scope":"该版本综合榜全部已评测模型","score":62.52,"source_id":"vals","source_model_id":"Kimi K3.1 Pro","source_url":"https://www.vals.ai/benchmarks/vals_index","total":27,"unit":"%","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"}],"effort":null,"mapping_type":null,"listed_days_ago":3,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.02,"canonical_slug":"nvidia/nemotron-3.5-super-vl-20261015","catalog_created_at":"2026-10-15T15:04:22+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"nvidia/nemotron-3.5-super-vl","input_price":0.15,"max_output_tokens":235929,"name":"NVIDIA: Nemotron 3.5 Super VL","new_model":true,"open_weights":true,"output_price":0.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"nvidia","provider_label":"nvidia","release_date":"2026-10-15","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"nvidia/nemotron-3.5-super-vl-20261015","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":11,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.1,"canonical_slug":"openai/gpt-6.1-sol-pro-20260929","catalog_created_at":"2026-09-29T17:28:06+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-6.1-sol-pro","input_price":2.0,"max_output_tokens":128000,"name":"OpenAI: GPT-6.1 Sol Pro","new_model":true,"open_weights":false,"output_price":10.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2026-09-29","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-6.1-sol-pro-20260929","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":27,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.05,"canonical_slug":"reflection/beam-1.0-501b-a23b-20261016","catalog_created_at":"2026-10-16T14:21:39+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"reflection/beam-1.0-501b-a23b","input_price":0.25,"max_output_tokens":131072,"name":"Reflection: Beam 1.0 501B A23B","new_model":true,"open_weights":true,"output_price":1.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"reflection","provider_label":"reflection","release_date":"2026-10-05","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"reflection/beam-1.0-501b-a23b-20261016","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":21,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"scaledown/extract-20260928","catalog_created_at":"2026-10-13T12:09:11+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"scaledown/extract","input_price":0.04,"max_output_tokens":900000,"name":"ScaleDown: Extract","new_model":true,"open_weights":false,"output_price":0.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"scaledown","provider_label":"scaledown","release_date":"2026-10-13","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"scaledown/extract-20260928","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":13,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"scaledown/summarize-20260928","catalog_created_at":"2026-10-13T12:09:20+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"scaledown/summarize","input_price":0.04,"max_output_tokens":4000,"name":"ScaleDown: Summarize","new_model":true,"open_weights":false,"output_price":0.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"scaledown","provider_label":"scaledown","release_date":"2026-10-13","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"scaledown/summarize-20260928","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":13,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.03,"canonical_slug":"unbiased/pareto-26.10-20261015","catalog_created_at":"2026-10-15T14:13:40+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"unbiased/pareto-26.10","input_price":0.8,"max_output_tokens":131072,"name":"Pareto 26.10","new_model":true,"open_weights":false,"output_price":3.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"unbiased","provider_label":"unbiased","release_date":"2026-10-15","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"unbiased/pareto-26.10-20261015","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":11,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.25,"canonical_slug":"unbiased/pareto-26.9-20260917","catalog_created_at":"2026-10-13T22:28:57+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"unbiased/pareto-26.9","input_price":2.5,"max_output_tokens":131072,"name":"Pareto 26.9","new_model":true,"open_weights":false,"output_price":7.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"unbiased","provider_label":"unbiased","release_date":"2026-10-13","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"unbiased/pareto-26.9-20260917","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":13,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.75,"canonical_slug":"aion-labs/aion-3.5-20260923","catalog_created_at":"2026-09-23T13:42:41+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"aion-labs/aion-3.5","input_price":3.0,"max_output_tokens":32768,"name":"AionLabs: Aion 3.5","new_model":false,"open_weights":false,"output_price":6.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"aion-labs","provider_label":"aion-labs","release_date":"2026-09-23","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"aion-labs/aion-3.5-20260923","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":33,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.18,"canonical_slug":"aion-labs/aion-3.5-mini-20260923","catalog_created_at":"2026-09-23T13:42:42+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"aion-labs/aion-3.5-mini","input_price":0.7,"max_output_tokens":32768,"name":"AionLabs: Aion 3.5 Mini","new_model":false,"open_weights":false,"output_price":1.4,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"aion-labs","provider_label":"aion-labs","release_date":"2026-09-23","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"aion-labs/aion-3.5-mini-20260923","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":33,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.2,"canonical_slug":"aion-labs/aion-2.0-20260223","catalog_created_at":"2026-02-23T21:15:06+00:00","catalog_status":"active","context_length":131072,"eligible":false,"extra_pricing":{},"id":"aion-labs/aion-2.0","input_price":0.8,"max_output_tokens":32768,"name":"AionLabs: Aion-2.0","new_model":false,"open_weights":false,"output_price":1.6,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"aion-labs","provider_label":"aion-labs","release_date":"2025-12-21","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"aion-labs/aion-2.0-20260223","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":309,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.75,"canonical_slug":"aion-labs/aion-3.0-20260707","catalog_created_at":"2026-07-07T16:51:35+00:00","catalog_status":"active","context_length":131072,"eligible":false,"extra_pricing":{},"id":"aion-labs/aion-3.0","input_price":3.0,"max_output_tokens":32768,"name":"AionLabs: Aion-3.0","new_model":false,"open_weights":false,"output_price":6.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"aion-labs","provider_label":"aion-labs","release_date":"2026-05-05","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"aion-labs/aion-3.0-20260707","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":174,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.18,"canonical_slug":"aion-labs/aion-3.0-mini-20260707","catalog_created_at":"2026-07-07T16:51:36+00:00","catalog_status":"active","context_length":131072,"eligible":false,"extra_pricing":{},"id":"aion-labs/aion-3.0-mini","input_price":0.7,"max_output_tokens":32768,"name":"AionLabs: Aion-3.0-Mini","new_model":false,"open_weights":false,"output_price":1.4,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"aion-labs","provider_label":"aion-labs","release_date":"2026-05-14","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"aion-labs/aion-3.0-mini-20260707","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":165,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"aion-labs/aion-rp-llama-3.1-8b","catalog_created_at":"2025-02-04T19:18:38+00:00","catalog_status":"active","context_length":32768,"eligible":false,"extra_pricing":{},"id":"aion-labs/aion-rp-llama-3.1-8b","input_price":0.8,"max_output_tokens":29491,"name":"AionLabs: Aion-RP 1.0 (8B)","new_model":false,"open_weights":true,"output_price":1.6,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"aion-labs","provider_label":"aion-labs","release_date":"2025-02-04","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"aion-labs/aion-rp-llama-3.1-8b","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":629,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"amazon/nova-2-lite-v1","catalog_created_at":"2025-12-02T17:31:12+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"amazon/nova-2-lite-v1","input_price":0.3,"max_output_tokens":65535,"name":"Amazon: Nova 2 Lite","new_model":false,"open_weights":false,"output_price":2.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"amazon","provider_label":"amazon","release_date":"2025-12-02","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"amazon/nova-2-lite-v1","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":328,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"amazon/nova-lite-v1","catalog_created_at":"2024-12-05T22:22:43+00:00","catalog_status":"active","context_length":300000,"eligible":false,"extra_pricing":{},"id":"amazon/nova-lite-v1","input_price":0.06,"max_output_tokens":5120,"name":"Amazon: Nova Lite 1.0","new_model":false,"open_weights":false,"output_price":0.24,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"amazon","provider_label":"amazon","release_date":"2024-12-05","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"amazon/nova-lite-v1","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":690,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"amazon/nova-micro-v1","catalog_created_at":"2024-12-05T22:20:37+00:00","catalog_status":"active","context_length":128000,"eligible":false,"extra_pricing":{},"id":"amazon/nova-micro-v1","input_price":0.035,"max_output_tokens":5120,"name":"Amazon: Nova Micro 1.0","new_model":false,"open_weights":false,"output_price":0.14,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"amazon","provider_label":"amazon","release_date":"2024-12-05","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"amazon/nova-micro-v1","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":690,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.625,"canonical_slug":"amazon/nova-premier-v1","catalog_created_at":"2025-10-31T22:38:52+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"amazon/nova-premier-v1","input_price":2.5,"max_output_tokens":32000,"name":"Amazon: Nova Premier 1.0","new_model":false,"open_weights":false,"output_price":12.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"amazon","provider_label":"amazon","release_date":"2025-04-30","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"amazon/nova-premier-v1","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":544,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"amazon/nova-pro-v1","catalog_created_at":"2024-12-05T22:05:03+00:00","catalog_status":"active","context_length":300000,"eligible":false,"extra_pricing":{},"id":"amazon/nova-pro-v1","input_price":0.8,"max_output_tokens":5120,"name":"Amazon: Nova Pro 1.0","new_model":false,"open_weights":false,"output_price":3.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"amazon","provider_label":"amazon","release_date":"2024-12-05","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"amazon/nova-pro-v1","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":690,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"anthracite-org/magnum-v4-72b","catalog_created_at":"2024-10-22T00:00:00+00:00","catalog_status":"active","context_length":32768,"eligible":false,"extra_pricing":{},"id":"anthracite-org/magnum-v4-72b","input_price":2.5,"max_output_tokens":4096,"name":"Magnum v4 72B","new_model":false,"open_weights":true,"output_price":5.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthracite-org","provider_label":"anthracite-org","release_date":"2024-10-22","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthracite-org/magnum-v4-72b","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":734,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.5,"canonical_slug":"anthropic/claude-4.5-opus-20251124","catalog_created_at":"2025-11-24T18:56:20+00:00","catalog_status":"active","context_length":200000,"eligible":false,"extra_pricing":{},"id":"anthropic/claude-opus-4.5","input_price":5.0,"max_output_tokens":64000,"name":"Anthropic: Claude Opus 4.5","new_model":false,"open_weights":false,"output_price":25.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthropic","provider_label":"anthropic","release_date":"2025-11-24","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthropic/claude-4.5-opus-20251124","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":153.01,"ci_low":147.81,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-11-24"],"catalog_release_date":"2025-11-24","source_model_date":"2025-11-24","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"anthropic/claude-4.5-opus-20251124","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":58,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":150.11,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-11-24","source_model_id":"Claude Opus 4.5","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":336,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.06,"canonical_slug":"arcee-ai/trinity-large-thinking","catalog_created_at":"2026-04-01T15:45:18+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"arcee-ai/trinity-large-thinking","input_price":0.25,"max_output_tokens":80000,"name":"Arcee AI: Trinity Large Thinking","new_model":false,"open_weights":true,"output_price":0.8,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"arcee-ai","provider_label":"arcee-ai","release_date":"2026-04-01","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"arcee-ai/trinity-large-thinking","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":208,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.1,"canonical_slug":"bytedance/ui-tars-1.5-7b","catalog_created_at":"2025-07-22T17:24:16+00:00","catalog_status":"active","context_length":128000,"eligible":false,"extra_pricing":{},"id":"bytedance/ui-tars-1.5-7b","input_price":0.1,"max_output_tokens":2048,"name":"ByteDance: UI-TARS 7B ","new_model":false,"open_weights":true,"output_price":0.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"bytedance","provider_label":"bytedance","release_date":"2025-07-22","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"bytedance/ui-tars-1.5-7b","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":461,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"bytedance-seed/seed-1.6-20250625","catalog_created_at":"2025-12-23T15:49:57+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"bytedance-seed/seed-1.6","input_price":0.25,"max_output_tokens":32768,"name":"ByteDance Seed: Seed 1.6","new_model":false,"open_weights":false,"output_price":2.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"bytedance-seed","provider_label":"bytedance-seed","release_date":"2025-10-15","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"bytedance-seed/seed-1.6-20250625","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":376,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"bytedance-seed/seed-1.6-flash-20250625","catalog_created_at":"2025-12-23T15:50:11+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"bytedance-seed/seed-1.6-flash","input_price":0.075,"max_output_tokens":32768,"name":"ByteDance Seed: Seed 1.6 Flash","new_model":false,"open_weights":false,"output_price":0.3,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"bytedance-seed","provider_label":"bytedance-seed","release_date":"2025-08-28","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"bytedance-seed/seed-1.6-flash-20250625","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":424,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"bytedance-seed/seed-2-1-turbo-20260810","catalog_created_at":"2026-08-12T16:29:36+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"bytedance-seed/seed-2-1-turbo","input_price":0.5,"max_output_tokens":235929,"name":"ByteDance Seed: Seed 2.1 Turbo","new_model":false,"open_weights":false,"output_price":2.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"bytedance-seed","provider_label":"bytedance-seed","release_date":"2026-06-23","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"bytedance-seed/seed-2-1-turbo-20260810","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":125,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"bytedance-seed/seed-2.0-code-20260730","catalog_created_at":"2026-08-12T16:05:01+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"bytedance-seed/seed-2.0-code","input_price":0.5,"max_output_tokens":131072,"name":"ByteDance Seed: Seed-2.0-Code","new_model":false,"open_weights":false,"output_price":3.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"bytedance-seed","provider_label":"bytedance-seed","release_date":"2026-02-14","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"bytedance-seed/seed-2.0-code-20260730","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":254,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"bytedance-seed/seed-2.0-lite-20260309","catalog_created_at":"2026-03-10T15:40:31+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"bytedance-seed/seed-2.0-lite","input_price":0.25,"max_output_tokens":131072,"name":"ByteDance Seed: Seed-2.0-Lite","new_model":false,"open_weights":false,"output_price":2.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"bytedance-seed","provider_label":"bytedance-seed","release_date":"2026-02-14","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"bytedance-seed/seed-2.0-lite-20260309","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":254,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"bytedance-seed/seed-2.0-mini-20260224","catalog_created_at":"2026-02-26T18:38:27+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"bytedance-seed/seed-2.0-mini","input_price":0.1,"max_output_tokens":131072,"name":"ByteDance Seed: Seed-2.0-Mini","new_model":false,"open_weights":false,"output_price":0.4,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"bytedance-seed","provider_label":"bytedance-seed","release_date":"2026-02-14","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"bytedance-seed/seed-2.0-mini-20260224","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":254,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"venice/uncensored","catalog_created_at":"2025-07-09T21:02:46+00:00","catalog_status":"active","context_length":128000,"eligible":false,"extra_pricing":{},"id":"cognitivecomputations/dolphin-mistral-24b-venice-edition","input_price":0.2,"max_output_tokens":8192,"name":"Venice: Uncensored","new_model":false,"open_weights":true,"output_price":0.9,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"cognitivecomputations","provider_label":"cognitivecomputations","release_date":"2025-07-09","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"venice/uncensored","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":474,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.15,"canonical_slug":"cohere/command-a-plus-05-2026","catalog_created_at":"2026-09-22T18:48:16+00:00","catalog_status":"active","context_length":192000,"eligible":false,"extra_pricing":{},"id":"cohere/command-a-plus","input_price":0.3,"max_output_tokens":64000,"name":"Cohere: Command A+","new_model":false,"open_weights":true,"output_price":1.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"cohere","provider_label":"cohere","release_date":"2026-05-20","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"cohere/command-a-plus-05-2026","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":159,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"cohere/command-r-plus-08-2024","catalog_created_at":"2024-08-30T00:00:00+00:00","catalog_status":"active","context_length":128000,"eligible":false,"extra_pricing":{},"id":"cohere/command-r-plus-08-2024","input_price":2.5,"max_output_tokens":4000,"name":"Cohere: Command R+ (08-2024)","new_model":false,"open_weights":true,"output_price":10.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"cohere","provider_label":"cohere","release_date":"2024-08-30","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"cohere/command-r-plus-08-2024","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":787,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"cohere/command-r7b-12-2024","catalog_created_at":"2024-12-14T06:35:52+00:00","catalog_status":"active","context_length":128000,"eligible":false,"extra_pricing":{},"id":"cohere/command-r7b-12-2024","input_price":0.0375,"max_output_tokens":4000,"name":"Cohere: Command R7B (12-2024)","new_model":false,"open_weights":true,"output_price":0.15,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"cohere","provider_label":"cohere","release_date":"2024-12-13","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"cohere/command-r7b-12-2024","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":682,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"deepseek/deepseek-chat-v3","catalog_created_at":"2024-12-26T19:28:40+00:00","catalog_status":"active","context_length":163840,"eligible":false,"extra_pricing":{},"id":"deepseek/deepseek-chat","input_price":0.2574,"max_output_tokens":16000,"name":"DeepSeek: DeepSeek V3","new_model":false,"open_weights":true,"output_price":1.0287,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"deepseek","provider_label":"deepseek","release_date":"2025-12-01","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"deepseek/deepseek-chat-v3","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":329,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.11,"canonical_slug":"deepseek/deepseek-chat-v3-0324","catalog_created_at":"2025-03-24T13:59:15+00:00","catalog_status":"active","context_length":163840,"eligible":false,"extra_pricing":{},"id":"deepseek/deepseek-chat-v3-0324","input_price":0.29,"max_output_tokens":115200,"name":"DeepSeek: DeepSeek V3 0324","new_model":false,"open_weights":true,"output_price":1.14,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"deepseek","provider_label":"deepseek","release_date":"2025-03-24","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"deepseek/deepseek-chat-v3-0324","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":581,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.13,"canonical_slug":"deepseek/deepseek-chat-v3.1","catalog_created_at":"2025-08-21T12:33:48+00:00","catalog_status":"active","context_length":163840,"eligible":false,"extra_pricing":{},"id":"deepseek/deepseek-chat-v3.1","input_price":0.25,"max_output_tokens":32768,"name":"DeepSeek: DeepSeek V3.1","new_model":false,"open_weights":true,"output_price":0.95,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"deepseek","provider_label":"deepseek","release_date":"2025-08-21","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"deepseek/deepseek-chat-v3.1","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":143.26,"ci_low":135.65,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":[],"catalog_release_date":"2025-08-21","source_model_date":"2025-08-21","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"deepseek/deepseek-chat-v3.1","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":123,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":139.79,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-08-21","source_model_id":"DeepSeek-V3.1","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":431,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.001414,"canonical_slug":"deepseek/deepseek-v4-flash-20260731","catalog_created_at":"2026-07-31T06:21:48+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"deepseek/deepseek-v4-flash-0731","input_price":0.04444,"max_output_tokens":131072,"name":"DeepSeek: DeepSeek V4 Flash 0731","new_model":false,"open_weights":true,"output_price":0.13332,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"deepseek","provider_label":"deepseek","release_date":"2026-07-31","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"deepseek/deepseek-v4-flash-20260731","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":156.86,"ci_low":152.06,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-07-31"],"catalog_release_date":"2026-07-31","source_model_date":"2026-07-31","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"deepseek/deepseek-v4-flash-20260731","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":40,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":154.76,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-07-31","source_model_id":"DeepSeek V4 Flash 0731","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":87,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.00686,"canonical_slug":"deepseek/deepseek-v4-flash-vision-exp-20260821","catalog_created_at":"2026-08-21T11:26:03+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"deepseek/deepseek-v4-flash-vision-exp","input_price":0.2156,"max_output_tokens":262144,"name":"DeepSeek: DeepSeek V4 Flash Vision Exp","new_model":false,"open_weights":true,"output_price":0.6468,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"deepseek","provider_label":"deepseek","release_date":"2026-08-21","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"deepseek/deepseek-v4-flash-vision-exp-20260821","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":66,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.3,"canonical_slug":"fireworks/ember-1-20260923","catalog_created_at":"2026-09-24T00:07:41+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"fireworks/ember-1","input_price":3.0,"max_output_tokens":943718,"name":"Fireworks: Ember-1","new_model":false,"open_weights":false,"output_price":15.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"fireworks","provider_label":"fireworks","release_date":"2026-09-22","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"fireworks/ember-1-20260923","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":34,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.01,"canonical_slug":"google/gemini-2.5-flash-lite","catalog_created_at":"2025-07-22T16:04:36+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"google/gemini-2.5-flash-lite","input_price":0.1,"max_output_tokens":65535,"name":"Google: Gemini 2.5 Flash Lite","new_model":false,"open_weights":false,"output_price":0.4,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2025-06-17","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-2.5-flash-lite","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":496,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.125,"canonical_slug":"google/gemini-2.5-pro-preview-06-05","catalog_created_at":"2025-06-05T15:27:37+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"google/gemini-2.5-pro-preview","input_price":1.25,"max_output_tokens":65536,"name":"Google: Gemini 2.5 Pro Preview 06-05","new_model":false,"open_weights":false,"output_price":10.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2025-06-05","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-2.5-pro-preview-06-05","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":508,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.05,"canonical_slug":"google/gemini-3-flash-preview-20251217","catalog_created_at":"2025-12-17T15:57:58+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"google/gemini-3-flash-preview","input_price":0.5,"max_output_tokens":65536,"name":"Google: Gemini 3 Flash Preview","new_model":false,"open_weights":false,"output_price":3.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2025-12-17","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-3-flash-preview-20251217","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":313,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.025,"canonical_slug":"google/gemini-3.1-flash-lite-20260507","catalog_created_at":"2026-05-07T15:47:08+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"google/gemini-3.1-flash-lite","input_price":0.25,"max_output_tokens":65536,"name":"Google: Gemini 3.1 Flash Lite","new_model":false,"open_weights":false,"output_price":1.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2026-05-07","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-3.1-flash-lite-20260507","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":172,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.2,"canonical_slug":"google/gemini-3.1-pro-preview-customtools-20260219","catalog_created_at":"2026-02-25T18:58:43+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"google/gemini-3.1-pro-preview-customtools","input_price":2.0,"max_output_tokens":65536,"name":"Google: Gemini 3.1 Pro Preview Custom Tools","new_model":false,"open_weights":false,"output_price":12.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2026-02-19","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-3.1-pro-preview-customtools-20260219","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":249,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.0375,"canonical_slug":"google/gemma-4-26b-a4b-it-20260403","catalog_created_at":"2026-04-03T14:53:09+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"google/gemma-4-26b-a4b-it","input_price":0.0675,"max_output_tokens":235929,"name":"Google: Gemma 4 26B A4B ","new_model":false,"open_weights":true,"output_price":0.225,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2026-04-02","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemma-4-26b-a4b-it-20260403","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":207,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"google/gemma-4-31b-it-20260402","catalog_created_at":"2026-04-02T16:48:06+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"google/gemma-4-31b-it","input_price":0.2,"max_output_tokens":235929,"name":"Google: Gemma 4 31B","new_model":false,"open_weights":true,"output_price":0.4,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2026-04-02","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemma-4-31b-it-20260402","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":207,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"google/lyria-3-clip-preview-20260330","catalog_created_at":"2026-03-30T21:47:35+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"google/lyria-3-clip-preview","input_price":0.0,"max_output_tokens":65536,"name":"Google: Lyria 3 Clip Preview","new_model":false,"open_weights":false,"output_price":0.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2026-03-25","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/lyria-3-clip-preview-20260330","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":215,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"google/lyria-3-pro-preview-20260330","catalog_created_at":"2026-03-30T21:48:06+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"google/lyria-3-pro-preview","input_price":0.0,"max_output_tokens":65536,"name":"Google: Lyria 3 Pro Preview","new_model":false,"open_weights":false,"output_price":0.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2026-03-25","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/lyria-3-pro-preview-20260330","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":215,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.03,"canonical_slug":"google/gemini-2.5-flash-image","catalog_created_at":"2025-10-07T20:53:51+00:00","catalog_status":"active","context_length":32768,"eligible":false,"extra_pricing":{},"id":"google/gemini-2.5-flash-image","input_price":0.3,"max_output_tokens":8192,"name":"Google: Nano Banana (Gemini 2.5 Flash Image)","new_model":false,"open_weights":false,"output_price":2.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2025-08-26","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-2.5-flash-image","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":426,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"google/gemini-3.1-flash-image-preview-20260226","catalog_created_at":"2026-02-26T15:25:58+00:00","catalog_status":"active","context_length":65536,"eligible":false,"extra_pricing":{},"id":"google/gemini-3.1-flash-image-preview","input_price":0.5,"max_output_tokens":58982,"name":"Google: Nano Banana 2 (Gemini 3.1 Flash Image Preview)","new_model":false,"open_weights":false,"output_price":3.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2026-02-26","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-3.1-flash-image-preview-20260226","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":242,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"google/gemini-3.1-flash-image-20260528","catalog_created_at":"2026-06-18T03:41:05+00:00","catalog_status":"active","context_length":131072,"eligible":false,"extra_pricing":{},"id":"google/gemini-3.1-flash-image","input_price":0.5,"max_output_tokens":32768,"name":"Google: Nano Banana 2 (Gemini 3.1 Flash Image)","new_model":false,"open_weights":false,"output_price":3.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2026-05-28","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-3.1-flash-image-20260528","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":151,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"google/gemini-3.1-flash-lite-image-20260630","catalog_created_at":"2026-06-30T16:33:45+00:00","catalog_status":"active","context_length":65536,"eligible":false,"extra_pricing":{},"id":"google/gemini-3.1-flash-lite-image","input_price":0.25,"max_output_tokens":58982,"name":"Google: Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image)","new_model":false,"open_weights":false,"output_price":1.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2026-06-30","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-3.1-flash-lite-image-20260630","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":118,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.2,"canonical_slug":"google/gemini-3-pro-image-preview-20251120","catalog_created_at":"2025-11-20T15:49:57+00:00","catalog_status":"active","context_length":65536,"eligible":false,"extra_pricing":{},"id":"google/gemini-3-pro-image-preview","input_price":2.0,"max_output_tokens":32768,"name":"Google: Nano Banana Pro (Gemini 3 Pro Image Preview)","new_model":false,"open_weights":false,"output_price":12.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2025-11-20","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-3-pro-image-preview-20251120","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":340,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.2,"canonical_slug":"google/gemini-3-pro-image-20260528","catalog_created_at":"2026-06-18T03:40:54+00:00","catalog_status":"active","context_length":131072,"eligible":false,"extra_pricing":{},"id":"google/gemini-3-pro-image","input_price":2.0,"max_output_tokens":32768,"name":"Google: Nano Banana Pro (Gemini 3 Pro Image)","new_model":false,"open_weights":false,"output_price":12.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2026-05-28","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-3-pro-image-20260528","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":151,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"gryphe/mythomax-l2-13b","catalog_created_at":"2023-07-02T00:00:00+00:00","catalog_status":"active","context_length":8192,"eligible":false,"extra_pricing":{},"id":"gryphe/mythomax-l2-13b","input_price":0.08,"max_output_tokens":3686,"name":"MythoMax 13B","new_model":false,"open_weights":true,"output_price":0.11,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"gryphe","provider_label":"gryphe","release_date":"2023-08-10","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"gryphe/mythomax-l2-13b","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":1173,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"ibm-granite/granite-4.0-h-micro","catalog_created_at":"2025-10-20T02:34:55+00:00","catalog_status":"active","context_length":131000,"eligible":false,"extra_pricing":{},"id":"ibm-granite/granite-4.0-h-micro","input_price":0.017,"max_output_tokens":117900,"name":"IBM: Granite 4.0 Micro","new_model":false,"open_weights":true,"output_price":0.112,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"ibm-granite","provider_label":"ibm-granite","release_date":"2025-10-20","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"ibm-granite/granite-4.0-h-micro","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":371,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.004,"canonical_slug":"inception/mercury-2.5-20260908","catalog_created_at":"2026-09-08T18:28:57+00:00","catalog_status":"active","context_length":260000,"eligible":false,"extra_pricing":{},"id":"inception/mercury-2.5","input_price":0.04,"max_output_tokens":65536,"name":"Inception: Mercury 2.5","new_model":false,"open_weights":false,"output_price":0.15,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"inception","provider_label":"inception","release_date":"2026-09-08","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"inception/mercury-2.5-20260908","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":48,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.0042,"canonical_slug":"inclusionai/ling-3.0-flash-20260723","catalog_created_at":"2026-07-23T14:56:20+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"inclusionai/ling-3.0-flash","input_price":0.021,"max_output_tokens":32768,"name":"inclusionAI: Ling 3.0 Flash","new_model":false,"open_weights":true,"output_price":0.063,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"inclusionai","provider_label":"inclusionai","release_date":"2026-07-23","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"inclusionai/ling-3.0-flash-20260723","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":95,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.012,"canonical_slug":"inclusionai/ling-3.0-flash-vl-20260910","catalog_created_at":"2026-09-10T16:01:54+00:00","catalog_status":"active","context_length":131072,"eligible":false,"extra_pricing":{},"id":"inclusionai/ling-3.0-flash-vl","input_price":0.06,"max_output_tokens":32768,"name":"inclusionAI: Ling 3.0 Flash VL","new_model":false,"open_weights":true,"output_price":0.18,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"inclusionai","provider_label":"inclusionai","release_date":"2026-09-10","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"inclusionai/ling-3.0-flash-vl-20260910","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":46,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.05,"canonical_slug":"inference-net/schematron-v2-small-20260902","catalog_created_at":"2026-09-12T01:35:33+00:00","catalog_status":"active","context_length":128000,"eligible":false,"extra_pricing":{},"id":"inference-net/schematron-v2-small","input_price":0.05,"max_output_tokens":4096,"name":"Inference.net: Schematron V2 Small","new_model":false,"open_weights":false,"output_price":0.23,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"inference-net","provider_label":"inference-net","release_date":"2026-04-16","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"inference-net/schematron-v2-small-20260902","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":193,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.03,"canonical_slug":"inference-net/schematron-v2-turbo-20260902","catalog_created_at":"2026-09-12T01:35:49+00:00","catalog_status":"active","context_length":128000,"eligible":false,"extra_pricing":{},"id":"inference-net/schematron-v2-turbo","input_price":0.03,"max_output_tokens":8192,"name":"Inference.net: Schematron V2 Turbo","new_model":false,"open_weights":false,"output_price":0.15,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"inference-net","provider_label":"inference-net","release_date":"2026-04-16","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"inference-net/schematron-v2-turbo-20260902","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":193,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.04,"canonical_slug":"liquid/lfm-2.5-2.6b-20260811","catalog_created_at":"2026-08-11T17:48:39+00:00","catalog_status":"active","context_length":128000,"eligible":false,"extra_pricing":{},"id":"liquid/lfm-2.5-2.6b","input_price":0.35,"max_output_tokens":8192,"name":"LiquidAI: LFM2.5-2.6B","new_model":false,"open_weights":true,"output_price":0.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"liquid","provider_label":"liquid","release_date":"2026-08-11","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"liquid/lfm-2.5-2.6b-20260811","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":76,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"mancer/weaver","catalog_created_at":"2023-08-02T00:00:00+00:00","catalog_status":"active","context_length":8000,"eligible":false,"extra_pricing":{},"id":"mancer/weaver","input_price":0.4,"max_output_tokens":6000,"name":"Mancer: Weaver (alpha)","new_model":false,"open_weights":false,"output_price":0.75,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"mancer","provider_label":"mancer","release_date":"2023-08-02","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"mancer/weaver","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":1181,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.006,"canonical_slug":"meituan/longcat-2.0-20260720","catalog_created_at":"2026-07-20T13:37:38+00:00","catalog_status":"active","context_length":1048756,"eligible":false,"extra_pricing":{},"id":"meituan/longcat-2.0","input_price":0.3,"max_output_tokens":262144,"name":"Meituan: LongCat 2.0","new_model":false,"open_weights":true,"output_price":1.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"meituan","provider_label":"meituan","release_date":"2026-06-30","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"meituan/longcat-2.0-20260720","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":118,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.04,"canonical_slug":"meta/muse-glimmer-30b-20260810","catalog_created_at":"2026-08-09T19:06:34+00:00","catalog_status":"active","context_length":131072,"eligible":false,"extra_pricing":{},"id":"meta/muse-glimmer-30b","input_price":0.3,"max_output_tokens":16384,"name":"Meta: Muse Glimmer 30B","new_model":false,"open_weights":true,"output_price":1.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"meta","provider_label":"meta","release_date":"2026-08-10","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"meta/muse-glimmer-30b-20260810","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":77,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.002,"canonical_slug":"meta/muse-spark-1.2-contributor-20260805","catalog_created_at":"2026-08-21T18:21:16+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"meta/muse-spark-1.2-contributor","input_price":0.1,"max_output_tokens":943718,"name":"Meta: Muse Spark 1.2 Contributor","new_model":false,"open_weights":false,"output_price":0.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"meta","provider_label":"meta","release_date":"2026-08-21","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"meta/muse-spark-1.2-contributor-20260805","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":66,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.002,"canonical_slug":"meta/muse-spark-1.3-contributor-20260902","catalog_created_at":"2026-09-02T20:38:39+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"meta/muse-spark-1.3-contributor","input_price":0.1,"max_output_tokens":943718,"name":"Meta: Muse Spark 1.3 Contributor","new_model":false,"open_weights":false,"output_price":0.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"meta","provider_label":"meta","release_date":"2026-09-02","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"meta/muse-spark-1.3-contributor-20260902","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":54,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"meta-llama/llama-guard-4-12b","catalog_created_at":"2025-04-30T01:06:33+00:00","catalog_status":"active","context_length":163840,"eligible":false,"extra_pricing":{},"id":"meta-llama/llama-guard-4-12b","input_price":0.18,"max_output_tokens":16384,"name":"Meta: Llama Guard 4 12B","new_model":false,"open_weights":true,"output_price":0.18,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"meta-llama","provider_label":"meta-llama","release_date":"2025-04-29","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"meta-llama/llama-guard-4-12b","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":545,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.03,"canonical_slug":"minimax/minimax-m2-her-20260123","catalog_created_at":"2026-01-23T14:07:19+00:00","catalog_status":"active","context_length":65536,"eligible":false,"extra_pricing":{},"id":"minimax/minimax-m2-her","input_price":0.3,"max_output_tokens":2048,"name":"MiniMax: MiniMax M2-her","new_model":false,"open_weights":false,"output_price":1.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"minimax","provider_label":"minimax","release_date":"2026-01-23","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"minimax/minimax-m2-her-20260123","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":276,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.03,"canonical_slug":"minimax/minimax-m2.1","catalog_created_at":"2025-12-23T01:56:37+00:00","catalog_status":"active","context_length":204800,"eligible":false,"extra_pricing":{},"id":"minimax/minimax-m2.1","input_price":0.3,"max_output_tokens":131072,"name":"MiniMax: MiniMax M2.1","new_model":false,"open_weights":true,"output_price":1.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"minimax","provider_label":"minimax","release_date":"2025-12-23","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"minimax/minimax-m2.1","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":307,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"minimax/minimax-01","catalog_created_at":"2025-01-15T04:31:02+00:00","catalog_status":"active","context_length":1000192,"eligible":false,"extra_pricing":{},"id":"minimax/minimax-01","input_price":0.2,"max_output_tokens":900172,"name":"MiniMax: MiniMax-01","new_model":false,"open_weights":true,"output_price":1.1,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"minimax","provider_label":"minimax","release_date":"2025-01-15","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"minimax/minimax-01","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":649,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.2,"canonical_slug":"mistralai/mistral-large","catalog_created_at":"2024-02-26T00:00:00+00:00","catalog_status":"active","context_length":128000,"eligible":false,"extra_pricing":{},"id":"mistralai/mistral-large","input_price":2.0,"max_output_tokens":102400,"name":"Mistral Large","new_model":false,"open_weights":false,"output_price":6.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"mistralai","provider_label":"mistralai","release_date":"2024-02-26","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"mistralai/mistral-large","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":124.74,"ci_low":114.75,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":[],"catalog_release_date":"2024-02-26","source_model_date":"2024-02-26","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"mistralai/mistral-large","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":188,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":122.12,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2024-02-26","source_model_id":"Mistral Large","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":973,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.2,"canonical_slug":"mistralai/mistral-large-2407","catalog_created_at":"2024-11-19T01:06:55+00:00","catalog_status":"active","context_length":131072,"eligible":false,"extra_pricing":{},"id":"mistralai/mistral-large-2407","input_price":2.0,"max_output_tokens":104857,"name":"Mistral Large 2407","new_model":false,"open_weights":true,"output_price":6.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"mistralai","provider_label":"mistralai","release_date":"2024-07-24","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"mistralai/mistral-large-2407","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":824,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.03,"canonical_slug":"mistralai/codestral-2508","catalog_created_at":"2025-08-01T20:20:30+00:00","catalog_status":"active","context_length":256000,"eligible":false,"extra_pricing":{},"id":"mistralai/codestral-2508","input_price":0.3,"max_output_tokens":204800,"name":"Mistral: Codestral 2508","new_model":false,"open_weights":false,"output_price":0.9,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"mistralai","provider_label":"mistralai","release_date":"2025-07-30","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"mistralai/codestral-2508","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":453,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.04,"canonical_slug":"mistralai/devstral-2512","catalog_created_at":"2025-12-09T13:03:39+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"mistralai/devstral-2512","input_price":0.4,"max_output_tokens":209715,"name":"Mistral: Devstral 2 2512","new_model":false,"open_weights":true,"output_price":2.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"mistralai","provider_label":"mistralai","release_date":"2025-12-09","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"mistralai/devstral-2512","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":321,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.02,"canonical_slug":"mistralai/ministral-14b-2512","catalog_created_at":"2025-12-02T13:22:15+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"mistralai/ministral-14b-2512","input_price":0.2,"max_output_tokens":209715,"name":"Mistral: Ministral 3 14B 2512","new_model":false,"open_weights":true,"output_price":0.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"mistralai","provider_label":"mistralai","release_date":"2025-12-02","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"mistralai/ministral-14b-2512","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":328,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.01,"canonical_slug":"mistralai/ministral-3b-2512","catalog_created_at":"2025-12-02T13:19:20+00:00","catalog_status":"active","context_length":131072,"eligible":false,"extra_pricing":{},"id":"mistralai/ministral-3b-2512","input_price":0.1,"max_output_tokens":104857,"name":"Mistral: Ministral 3 3B 2512","new_model":false,"open_weights":true,"output_price":0.1,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"mistralai","provider_label":"mistralai","release_date":"2025-12-02","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"mistralai/ministral-3b-2512","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":328,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.015,"canonical_slug":"mistralai/ministral-8b-2512","catalog_created_at":"2025-12-02T13:20:54+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"mistralai/ministral-8b-2512","input_price":0.15,"max_output_tokens":209715,"name":"Mistral: Ministral 3 8B 2512","new_model":false,"open_weights":true,"output_price":0.15,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"mistralai","provider_label":"mistralai","release_date":"2025-12-02","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"mistralai/ministral-8b-2512","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":328,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.05,"canonical_slug":"mistralai/mistral-large-2512","catalog_created_at":"2025-12-01T21:27:52+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"mistralai/mistral-large-2512","input_price":0.5,"max_output_tokens":209715,"name":"Mistral: Mistral Large 3 2512","new_model":false,"open_weights":true,"output_price":1.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"mistralai","provider_label":"mistralai","release_date":"2025-12-02","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"mistralai/mistral-large-2512","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":328,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"mistralai/mistral-nemo","catalog_created_at":"2024-07-19T00:00:00+00:00","catalog_status":"active","context_length":131072,"eligible":false,"extra_pricing":{},"id":"mistralai/mistral-nemo","input_price":0.029,"max_output_tokens":16384,"name":"Mistral: Mistral Nemo","new_model":false,"open_weights":true,"output_price":0.03,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"mistralai","provider_label":"mistralai","release_date":"2024-07-18","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"mistralai/mistral-nemo","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":121.37,"ci_low":111.1,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":[],"catalog_release_date":"2024-07-18","source_model_date":"2024-07-18","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"mistralai/mistral-nemo","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":201,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":118.65,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2024-07-18","source_model_id":"Mistral NeMo","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":830,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"mistralai/mistral-small-3.2-24b-instruct-2506","catalog_created_at":"2025-06-20T18:10:16+00:00","catalog_status":"active","context_length":256000,"eligible":false,"extra_pricing":{},"id":"mistralai/mistral-small-3.2-24b-instruct","input_price":0.075,"max_output_tokens":16384,"name":"Mistral: Mistral Small 3.2 24B","new_model":false,"open_weights":true,"output_price":0.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"mistralai","provider_label":"mistralai","release_date":"2025-06-20","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"mistralai/mistral-small-3.2-24b-instruct-2506","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":493,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.015,"canonical_slug":"mistralai/mistral-small-2603","catalog_created_at":"2026-03-16T21:14:45+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"mistralai/mistral-small-2603","input_price":0.15,"max_output_tokens":209715,"name":"Mistral: Mistral Small 4","new_model":false,"open_weights":true,"output_price":0.6,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"mistralai","provider_label":"mistralai","release_date":"2026-03-16","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"mistralai/mistral-small-2603","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":224,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.2,"canonical_slug":"mistralai/mixtral-8x22b-instruct","catalog_created_at":"2024-04-17T00:00:00+00:00","catalog_status":"active","context_length":65536,"eligible":false,"extra_pricing":{},"id":"mistralai/mixtral-8x22b-instruct","input_price":2.0,"max_output_tokens":52428,"name":"Mistral: Mixtral 8x22B Instruct","new_model":false,"open_weights":true,"output_price":6.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"mistralai","provider_label":"mistralai","release_date":"2024-04-17","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"mistralai/mixtral-8x22b-instruct","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":922,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.02,"canonical_slug":"mistralai/mistral-saba-2502","catalog_created_at":"2025-02-17T14:40:39+00:00","catalog_status":"active","context_length":32768,"eligible":false,"extra_pricing":{},"id":"mistralai/mistral-saba","input_price":0.2,"max_output_tokens":26214,"name":"Mistral: Saba","new_model":false,"open_weights":false,"output_price":0.6,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"mistralai","provider_label":"mistralai","release_date":"2025-02-17","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"mistralai/mistral-saba-2502","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":616,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.01,"canonical_slug":"mistralai/voxtral-small-24b-2507","catalog_created_at":"2025-10-30T14:39:04+00:00","catalog_status":"active","context_length":32768,"eligible":false,"extra_pricing":{},"id":"mistralai/voxtral-small-24b-2507","input_price":0.1,"max_output_tokens":26214,"name":"Mistral: Voxtral Small 24B 2507","new_model":false,"open_weights":true,"output_price":0.3,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"mistralai","provider_label":"mistralai","release_date":"2025-07-15","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"mistralai/voxtral-small-24b-2507","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":468,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"moonshotai/kimi-k2","catalog_created_at":"2025-07-11T19:47:32+00:00","catalog_status":"active","context_length":131072,"eligible":false,"extra_pricing":{},"id":"moonshotai/kimi-k2","input_price":0.57,"max_output_tokens":98304,"name":"MoonshotAI: Kimi K2 0711","new_model":false,"open_weights":true,"output_price":2.3,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"moonshotai","provider_label":"moonshotai","release_date":"2025-07-11","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"moonshotai/kimi-k2","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":472,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"moonshotai/kimi-k2-0905","catalog_created_at":"2025-09-04T21:25:47+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"moonshotai/kimi-k2-0905","input_price":0.6,"max_output_tokens":98304,"name":"MoonshotAI: Kimi K2 0905","new_model":false,"open_weights":true,"output_price":2.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"moonshotai","provider_label":"moonshotai","release_date":"2025-09-05","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"moonshotai/kimi-k2-0905","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":416,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.15,"canonical_slug":"moonshotai/kimi-k2-thinking-20251106","catalog_created_at":"2025-11-06T14:50:22+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"moonshotai/kimi-k2-thinking","input_price":0.6,"max_output_tokens":98304,"name":"MoonshotAI: Kimi K2 Thinking","new_model":false,"open_weights":true,"output_price":2.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"moonshotai","provider_label":"moonshotai","release_date":"2025-11-06","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"moonshotai/kimi-k2-thinking-20251106","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":147.51,"ci_low":143.22,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-11-06"],"catalog_release_date":"2025-11-06","source_model_date":"2025-11-06","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"moonshotai/kimi-k2-thinking-20251106","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":85,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":145.82,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-11-06","source_model_id":"Kimi K2 Thinking","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":354,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.15,"canonical_slug":"moonshotai/kimi-k2.7-code-20260612","catalog_created_at":"2026-06-12T12:12:41+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"moonshotai/kimi-k2.7-code","input_price":0.6712,"max_output_tokens":235929,"name":"MoonshotAI: Kimi K2.7 Code","new_model":false,"open_weights":true,"output_price":3.35,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"moonshotai","provider_label":"moonshotai","release_date":"2026-06-12","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"moonshotai/kimi-k2.7-code-20260612","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":151.82,"ci_low":148.17,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-06-12"],"catalog_release_date":"2026-06-12","source_model_date":"2026-06-12","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"moonshotai/kimi-k2.7-code-20260612","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":58,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":150.11,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-06-12","source_model_id":"Kimi K2.7 Code","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"}],"effort":null,"mapping_type":null,"listed_days_ago":136,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"morph/morph-v3-fast","catalog_created_at":"2025-07-07T17:40:02+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"morph/morph-v3-fast","input_price":0.8,"max_output_tokens":131072,"name":"Morph: Morph V3 Fast","new_model":false,"open_weights":false,"output_price":1.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"morph","provider_label":"morph","release_date":"2025-07-07","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"morph/morph-v3-fast","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":476,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"morph/morph-v3-large","catalog_created_at":"2025-07-07T17:54:18+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"morph/morph-v3-large","input_price":0.9,"max_output_tokens":131072,"name":"Morph: Morph V3 Large","new_model":false,"open_weights":false,"output_price":1.9,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"morph","provider_label":"morph","release_date":"2025-07-07","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"morph/morph-v3-large","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":476,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.0025,"canonical_slug":"nex-agi/nex-n2.5-mini-20260908","catalog_created_at":"2026-09-08T17:54:21+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"nex-agi/nex-n2.5-mini","input_price":0.025,"max_output_tokens":235929,"name":"Nex AGI: Nex-N2.5-Mini","new_model":false,"open_weights":true,"output_price":0.1,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"nex-agi","provider_label":"nex-agi","release_date":"2026-09-08","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"nex-agi/nex-n2.5-mini-20260908","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":48,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.015,"canonical_slug":"nex-agi/nex-n2.5-pro-20260907","catalog_created_at":"2026-09-08T17:54:10+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"nex-agi/nex-n2.5-pro","input_price":0.075,"max_output_tokens":235929,"name":"Nex AGI: Nex-N2.5-Pro","new_model":false,"open_weights":true,"output_price":0.25,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"nex-agi","provider_label":"nex-agi","release_date":"2026-09-08","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"nex-agi/nex-n2.5-pro-20260907","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":48,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"nousresearch/hermes-3-llama-3.1-405b","catalog_created_at":"2024-08-16T00:00:00+00:00","catalog_status":"active","context_length":131072,"eligible":false,"extra_pricing":{},"id":"nousresearch/hermes-3-llama-3.1-405b","input_price":1.0,"max_output_tokens":16384,"name":"Nous: Hermes 3 405B Instruct","new_model":false,"open_weights":true,"output_price":1.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"nousresearch","provider_label":"nousresearch","release_date":"2024-08-16","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"nousresearch/hermes-3-llama-3.1-405b","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":801,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"nousresearch/hermes-3-llama-3.1-70b","catalog_created_at":"2024-08-18T00:00:00+00:00","catalog_status":"active","context_length":131072,"eligible":false,"extra_pricing":{},"id":"nousresearch/hermes-3-llama-3.1-70b","input_price":0.7,"max_output_tokens":16384,"name":"Nous: Hermes 3 70B Instruct","new_model":false,"open_weights":true,"output_price":0.7,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"nousresearch","provider_label":"nousresearch","release_date":"2024-08-18","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"nousresearch/hermes-3-llama-3.1-70b","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":799,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"nousresearch/hermes-4-405b","catalog_created_at":"2025-08-26T19:11:03+00:00","catalog_status":"active","context_length":131072,"eligible":false,"extra_pricing":{},"id":"nousresearch/hermes-4-405b","input_price":1.0,"max_output_tokens":117964,"name":"Nous: Hermes 4 405B","new_model":false,"open_weights":true,"output_price":3.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"nousresearch","provider_label":"nousresearch","release_date":"2025-08-26","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"nousresearch/hermes-4-405b","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":426,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"nvidia/nemotron-3-nano-30b-a3b","catalog_created_at":"2025-12-14T16:54:35+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"nvidia/nemotron-3-nano-30b-a3b","input_price":0.06,"max_output_tokens":235929,"name":"NVIDIA: Nemotron 3 Nano 30B A3B","new_model":false,"open_weights":true,"output_price":0.24,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"nvidia","provider_label":"nvidia","release_date":"2025-12-15","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"nvidia/nemotron-3-nano-30b-a3b","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":315,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.15,"canonical_slug":"nvidia/nemotron-3-super-120b-a12b-20230311","catalog_created_at":"2026-03-11T16:07:19+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"nvidia/nemotron-3-super-120b-a12b","input_price":0.3,"max_output_tokens":235929,"name":"NVIDIA: Nemotron 3 Super","new_model":false,"open_weights":true,"output_price":2.4,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"nvidia","provider_label":"nvidia","release_date":"2026-03-11","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"nvidia/nemotron-3-super-120b-a12b-20230311","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":229,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.12,"canonical_slug":"nvidia/nemotron-3-ultra-550b-a55b-20260604","catalog_created_at":"2026-06-04T05:33:28+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"nvidia/nemotron-3-ultra-550b-a55b","input_price":0.6,"max_output_tokens":182520,"name":"NVIDIA: Nemotron 3 Ultra","new_model":false,"open_weights":true,"output_price":2.4,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"nvidia","provider_label":"nvidia","release_date":"2026-06-04","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"nvidia/nemotron-3-ultra-550b-a55b-20260604","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":148.22,"ci_low":143.68,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-06-04"],"catalog_release_date":"2026-06-04","source_model_date":"2026-06-04","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"nvidia/nemotron-3-ultra-550b-a55b-20260604","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":84,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":145.97,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-06-04","source_model_id":"Nemotron 3 Ultra","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":144,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"nvidia/nemotron-3.5-content-safety-20260604","catalog_created_at":"2026-06-04T14:04:24+00:00","catalog_status":"active","context_length":131072,"eligible":false,"extra_pricing":{},"id":"nvidia/nemotron-3.5-content-safety","input_price":0.2,"max_output_tokens":117964,"name":"NVIDIA: Nemotron 3.5 Content Safety","new_model":false,"open_weights":true,"output_price":0.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"nvidia","provider_label":"nvidia","release_date":"2026-06-02","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"nvidia/nemotron-3.5-content-safety-20260604","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":146,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.04,"canonical_slug":"nvidia/nemotron-3.5-lightning-20260807","catalog_created_at":"2026-08-11T12:52:31+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"nvidia/nemotron-3.5-lightning","input_price":0.07,"max_output_tokens":235929,"name":"NVIDIA: Nemotron 3.5 Lightning","new_model":false,"open_weights":true,"output_price":0.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"nvidia","provider_label":"nvidia","release_date":"2026-08-11","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"nvidia/nemotron-3.5-lightning-20260807","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":76,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"nvidia/switchyard","catalog_created_at":"2026-09-21T21:25:47+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"nvidia/switchyard","input_price":null,"max_output_tokens":null,"name":"NVIDIA: Switchyard","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"nvidia","provider_label":"nvidia","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"nvidia/switchyard","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":35,"new_model_basis":"catalog_created_at","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-audio","catalog_created_at":"2026-01-19T22:42:49+00:00","catalog_status":"active","context_length":128000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-audio","input_price":2.5,"max_output_tokens":16384,"name":"OpenAI: GPT Audio","new_model":false,"open_weights":false,"output_price":10.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2025-08-28","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-audio","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":424,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-audio-mini","catalog_created_at":"2026-01-19T21:50:19+00:00","catalog_status":"active","context_length":128000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-audio-mini","input_price":0.6,"max_output_tokens":16384,"name":"OpenAI: GPT Audio Mini","new_model":false,"open_weights":false,"output_price":2.4,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2025-10-06","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-audio-mini","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":385,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.5,"canonical_slug":"openai/gpt-chat-latest-20260505","catalog_created_at":"2026-05-05T16:56:52+00:00","catalog_status":"active","context_length":400000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-chat-latest","input_price":5.0,"max_output_tokens":128000,"name":"OpenAI: GPT Chat Latest","new_model":false,"open_weights":false,"output_price":30.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2026-05-05","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-chat-latest-20260505","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":174,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-3.5-turbo-0613","catalog_created_at":"2024-01-25T00:00:00+00:00","catalog_status":"active","context_length":4095,"eligible":false,"extra_pricing":{},"id":"openai/gpt-3.5-turbo-0613","input_price":1.0,"max_output_tokens":3685,"name":"OpenAI: GPT-3.5 Turbo (older v0613)","new_model":false,"open_weights":false,"output_price":2.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2024-01-25","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-3.5-turbo-0613","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":1005,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-3.5-turbo-16k","catalog_created_at":"2023-08-28T00:00:00+00:00","catalog_status":"active","context_length":16385,"eligible":false,"extra_pricing":{},"id":"openai/gpt-3.5-turbo-16k","input_price":3.0,"max_output_tokens":4096,"name":"OpenAI: GPT-3.5 Turbo 16k","new_model":false,"open_weights":false,"output_price":4.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2023-08-28","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-3.5-turbo-16k","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":1155,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-4","catalog_created_at":"2023-05-28T00:00:00+00:00","catalog_status":"active","context_length":8191,"eligible":false,"extra_pricing":{},"id":"openai/gpt-4","input_price":30.0,"max_output_tokens":4096,"name":"OpenAI: GPT-4","new_model":false,"open_weights":false,"output_price":60.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2023-11-06","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-4","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":1085,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":1.25,"canonical_slug":"openai/gpt-4o","catalog_created_at":"2024-05-13T00:00:00+00:00","catalog_status":"active","context_length":128000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-4o","input_price":2.5,"max_output_tokens":16384,"name":"OpenAI: GPT-4o","new_model":false,"open_weights":false,"output_price":10.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2024-05-13","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-4o","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":896,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":1.25,"canonical_slug":"openai/gpt-4o-2024-11-20","catalog_created_at":"2024-11-20T18:33:14+00:00","catalog_status":"active","context_length":128000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-4o-2024-11-20","input_price":2.5,"max_output_tokens":16384,"name":"OpenAI: GPT-4o (2024-11-20)","new_model":false,"open_weights":false,"output_price":10.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2024-11-20","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-4o-2024-11-20","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":705,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.075,"canonical_slug":"openai/gpt-4o-mini","catalog_created_at":"2024-07-18T00:00:00+00:00","catalog_status":"active","context_length":128000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-4o-mini","input_price":0.15,"max_output_tokens":16384,"name":"OpenAI: GPT-4o-mini","new_model":false,"open_weights":false,"output_price":0.6,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2024-07-18","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-4o-mini","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":128.49,"ci_low":120.11,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":[],"catalog_release_date":"2024-07-18","source_model_date":"2024-07-18","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-4o-mini","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":177,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":126.46,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2024-07-18","source_model_id":"GPT-4o mini","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":830,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":1.25,"canonical_slug":"openai/gpt-5-image","catalog_created_at":"2025-10-14T13:19:46+00:00","catalog_status":"active","context_length":400000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5-image","input_price":10.0,"max_output_tokens":128000,"name":"OpenAI: GPT-5 Image","new_model":false,"open_weights":false,"output_price":10.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2025-10-14","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5-image","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":377,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.25,"canonical_slug":"openai/gpt-5-image-mini","catalog_created_at":"2025-10-16T14:23:03+00:00","catalog_status":"active","context_length":400000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5-image-mini","input_price":2.5,"max_output_tokens":128000,"name":"OpenAI: GPT-5 Image Mini","new_model":false,"open_weights":false,"output_price":2.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2025-10-16","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5-image-mini","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":375,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-5-pro-2025-10-06","catalog_created_at":"2025-10-06T18:51:03+00:00","catalog_status":"active","context_length":400000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5-pro","input_price":15.0,"max_output_tokens":128000,"name":"OpenAI: GPT-5 Pro","new_model":false,"open_weights":false,"output_price":120.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2025-10-06","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5-pro-2025-10-06","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":385,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.13,"canonical_slug":"openai/gpt-5.1-codex-20251113","catalog_created_at":"2025-11-13T18:58:18+00:00","catalog_status":"active","context_length":400000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5.1-codex","input_price":1.25,"max_output_tokens":128000,"name":"OpenAI: GPT-5.1-Codex","new_model":false,"open_weights":false,"output_price":10.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2025-11-13","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.1-codex-20251113","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"}],"effort":null,"mapping_type":null,"listed_days_ago":347,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.125,"canonical_slug":"openai/gpt-5.1-codex-max-20251204","catalog_created_at":"2025-12-04T20:08:54+00:00","catalog_status":"active","context_length":400000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5.1-codex-max","input_price":1.25,"max_output_tokens":128000,"name":"OpenAI: GPT-5.1-Codex-Max","new_model":false,"open_weights":false,"output_price":10.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2025-11-13","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.1-codex-max-20251204","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":347,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.03,"canonical_slug":"openai/gpt-5.1-codex-mini-20251113","catalog_created_at":"2025-11-13T18:17:00+00:00","catalog_status":"active","context_length":400000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5.1-codex-mini","input_price":0.25,"max_output_tokens":128000,"name":"OpenAI: GPT-5.1-Codex-Mini","new_model":false,"open_weights":false,"output_price":2.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2025-11-13","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.1-codex-mini-20251113","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"}],"effort":null,"mapping_type":null,"listed_days_ago":347,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.175,"canonical_slug":"openai/gpt-5.2-chat-20251211","catalog_created_at":"2025-12-10T18:03:03+00:00","catalog_status":"active","context_length":128000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5.2-chat","input_price":1.75,"max_output_tokens":32000,"name":"OpenAI: GPT-5.2 Chat","new_model":false,"open_weights":false,"output_price":14.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2025-12-11","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.2-chat-20251211","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":319,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-5.2-pro-20251211","catalog_created_at":"2025-12-10T18:03:00+00:00","catalog_status":"active","context_length":400000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5.2-pro","input_price":21.0,"max_output_tokens":128000,"name":"OpenAI: GPT-5.2 Pro","new_model":false,"open_weights":false,"output_price":168.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2025-12-11","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.2-pro-20251211","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":159.98,"ci_low":153.89,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-12-11"],"catalog_release_date":"2025-12-11","source_model_date":"2025-12-11","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-5.2-pro-20251211","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":30,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":156.33,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-12-11","source_model_id":"GPT-5.2 Pro","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":319,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.175,"canonical_slug":"openai/gpt-5.2-codex-20260114","catalog_created_at":"2026-01-14T16:48:35+00:00","catalog_status":"active","context_length":400000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5.2-codex","input_price":1.75,"max_output_tokens":128000,"name":"OpenAI: GPT-5.2-Codex","new_model":false,"open_weights":false,"output_price":14.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2025-12-11","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.2-codex-20260114","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"}],"effort":null,"mapping_type":null,"listed_days_ago":319,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.175,"canonical_slug":"openai/gpt-5.3-codex-20260224","catalog_created_at":"2026-02-24T18:52:44+00:00","catalog_status":"active","context_length":400000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5.3-codex","input_price":1.75,"max_output_tokens":128000,"name":"OpenAI: GPT-5.3-Codex","new_model":false,"open_weights":false,"output_price":14.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2026-02-05","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.3-codex-20260224","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":263,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":2.0,"canonical_slug":"openai/gpt-5.4-image-2-20260421","catalog_created_at":"2026-04-21T18:52:08+00:00","catalog_status":"active","context_length":272000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5.4-image-2","input_price":8.0,"max_output_tokens":128000,"name":"OpenAI: GPT-5.4 Image 2","new_model":false,"open_weights":false,"output_price":15.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2026-04-21","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.4-image-2-20260421","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":188,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-5.4-pro-20260305","catalog_created_at":"2026-03-05T18:12:46+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5.4-pro","input_price":30.0,"max_output_tokens":128000,"name":"OpenAI: GPT-5.4 Pro","new_model":false,"open_weights":false,"output_price":180.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2026-03-05","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.4-pro-20260305","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":162.7,"ci_low":157.1,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-03-05"],"catalog_release_date":"2026-03-05","source_model_date":"2026-03-05","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-5.4-pro-20260305","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":15,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":159.57,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-03-05","source_model_id":"GPT-5.4 Pro","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":235,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-5.5-pro-20260423","catalog_created_at":"2026-04-24T17:31:36+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5.5-pro","input_price":30.0,"max_output_tokens":128000,"name":"OpenAI: GPT-5.5 Pro","new_model":false,"open_weights":false,"output_price":180.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2026-04-23","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.5-pro-20260423","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":166.6,"ci_low":160.04,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-04-23"],"catalog_release_date":"2026-04-23","source_model_date":"2026-04-23","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-5.5-pro-20260423","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":10,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":162.91,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-04-23","source_model_id":"GPT-5.5 Pro","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":186,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.02,"canonical_slug":"openai/gpt-5.6-luna-pro-20260709","catalog_created_at":"2026-07-09T09:54:27+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5.6-luna-pro","input_price":0.2,"max_output_tokens":128000,"name":"OpenAI: GPT-5.6 Luna Pro","new_model":false,"open_weights":false,"output_price":1.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2026-07-09","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.6-luna-pro-20260709","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":109,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.4,"canonical_slug":"openai/gpt-5.6-sol-pro-20260709","catalog_created_at":"2026-07-09T09:54:14+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5.6-sol-pro","input_price":4.0,"max_output_tokens":128000,"name":"OpenAI: GPT-5.6 Sol Pro","new_model":false,"open_weights":false,"output_price":20.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2026-07-09","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.6-sol-pro-20260709","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":109,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.2,"canonical_slug":"openai/gpt-5.6-terra-pro-20260709","catalog_created_at":"2026-07-09T09:54:21+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5.6-terra-pro","input_price":2.0,"max_output_tokens":128000,"name":"OpenAI: GPT-5.6 Terra Pro","new_model":false,"open_weights":false,"output_price":12.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2026-07-09","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.6-terra-pro-20260709","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":109,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":1.0,"canonical_slug":"openai/gpt-6-astra-pro-20260903","catalog_created_at":"2026-09-04T20:13:55+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-6-astra-pro","input_price":10.0,"max_output_tokens":128000,"name":"OpenAI: GPT-6 Astra Pro","new_model":false,"open_weights":false,"output_price":50.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2026-09-04","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-6-astra-pro-20260903","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":52,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.01,"canonical_slug":"openai/gpt-6-luna-pro-20260922","catalog_created_at":"2026-09-22T18:13:11+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-6-luna-pro","input_price":0.1,"max_output_tokens":128000,"name":"OpenAI: GPT-6 Luna Pro","new_model":false,"open_weights":false,"output_price":0.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2026-09-22","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-6-luna-pro-20260922","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":34,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.2,"canonical_slug":"openai/gpt-6-sol-pro-20260922","catalog_created_at":"2026-09-22T18:13:01+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-6-sol-pro","input_price":2.0,"max_output_tokens":128000,"name":"OpenAI: GPT-6 Sol Pro","new_model":false,"open_weights":false,"output_price":10.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2026-09-22","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-6-sol-pro-20260922","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":34,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.0375,"canonical_slug":"openai/gpt-oss-safeguard-20b","catalog_created_at":"2025-10-29T15:47:16+00:00","catalog_status":"active","context_length":131072,"eligible":false,"extra_pricing":{},"id":"openai/gpt-oss-safeguard-20b","input_price":0.075,"max_output_tokens":65536,"name":"OpenAI: gpt-oss-safeguard-20b","new_model":false,"open_weights":true,"output_price":0.3,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2025-10-29","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-oss-safeguard-20b","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":362,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/o1-pro","catalog_created_at":"2025-03-19T22:26:51+00:00","catalog_status":"active","context_length":200000,"eligible":false,"extra_pricing":{},"id":"openai/o1-pro","input_price":150.0,"max_output_tokens":100000,"name":"OpenAI: o1-pro","new_model":false,"open_weights":false,"output_price":600.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2025-03-19","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/o1-pro","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":586,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/o3-pro-2025-06-10","catalog_created_at":"2025-06-10T23:32:32+00:00","catalog_status":"active","context_length":200000,"eligible":false,"extra_pricing":{},"id":"openai/o3-pro","input_price":20.0,"max_output_tokens":100000,"name":"OpenAI: o3 Pro","new_model":false,"open_weights":false,"output_price":80.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2025-06-10","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/o3-pro-2025-06-10","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":150.1,"ci_low":145.56,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-06-10"],"catalog_release_date":"2025-06-10","source_model_date":"2025-06-10","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/o3-pro-2025-06-10","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":70,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":147.33,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-06-10","source_model_id":"o3-pro","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":503,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.275,"canonical_slug":"openai/o4-mini-high-2025-04-16","catalog_created_at":"2025-04-16T17:23:32+00:00","catalog_status":"active","context_length":200000,"eligible":false,"extra_pricing":{},"id":"openai/o4-mini-high","input_price":1.1,"max_output_tokens":100000,"name":"OpenAI: o4 Mini High","new_model":false,"open_weights":false,"output_price":4.4,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":"2025-04-16","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/o4-mini-high-2025-04-16","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":558,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openrouter/auto","catalog_created_at":"2023-11-08T00:00:00+00:00","catalog_status":"active","context_length":2000000,"eligible":false,"extra_pricing":{},"id":"openrouter/auto","input_price":null,"max_output_tokens":null,"name":"Auto Router","new_model":false,"open_weights":false,"output_price":null,"price_source":null,"price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openrouter","provider_label":"openrouter","release_date":"2023-11-08","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openrouter/auto","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":1083,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openrouter/auto-beta","catalog_created_at":"2026-07-17T17:59:25+00:00","catalog_status":"active","context_length":2000000,"eligible":false,"extra_pricing":{},"id":"openrouter/auto-beta","input_price":null,"max_output_tokens":null,"name":"Auto Router (Beta)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openrouter","provider_label":"openrouter","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openrouter/auto-beta","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":101,"new_model_basis":"catalog_created_at","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openrouter/bodybuilder","catalog_created_at":"2025-12-05T03:00:53+00:00","catalog_status":"active","context_length":128000,"eligible":false,"extra_pricing":{},"id":"openrouter/bodybuilder","input_price":null,"max_output_tokens":null,"name":"Body Builder (beta)","new_model":false,"open_weights":false,"output_price":null,"price_source":null,"price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openrouter","provider_label":"openrouter","release_date":"2025-12-05","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openrouter/bodybuilder","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":325,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openrouter/free","catalog_created_at":"2026-02-01T03:43:47+00:00","catalog_status":"active","context_length":200000,"eligible":false,"extra_pricing":{},"id":"openrouter/free","input_price":0.0,"max_output_tokens":null,"name":"Free Models Router","new_model":false,"open_weights":false,"output_price":0.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openrouter","provider_label":"openrouter","release_date":"2026-02-01","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openrouter/free","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":267,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openrouter/fusion","catalog_created_at":"2026-06-13T17:27:27+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"openrouter/fusion","input_price":null,"max_output_tokens":null,"name":"OpenRouter: Fusion","new_model":false,"open_weights":false,"output_price":null,"price_source":null,"price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openrouter","provider_label":"openrouter","release_date":"2026-06-13","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openrouter/fusion","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":135,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openrouter/pareto-code","catalog_created_at":"2026-04-21T05:05:00+00:00","catalog_status":"active","context_length":2000000,"eligible":false,"extra_pricing":{},"id":"openrouter/pareto-code","input_price":null,"max_output_tokens":null,"name":"Pareto Code Router","new_model":false,"open_weights":false,"output_price":null,"price_source":null,"price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openrouter","provider_label":"openrouter","release_date":"2026-04-21","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openrouter/pareto-code","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":188,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"perceptron/perceptron-mk1-20260512","catalog_created_at":"2026-05-12T14:43:49+00:00","catalog_status":"active","context_length":32768,"eligible":false,"extra_pricing":{},"id":"perceptron/perceptron-mk1","input_price":0.15,"max_output_tokens":8192,"name":"Perceptron: Perceptron Mk1","new_model":false,"open_weights":false,"output_price":1.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"perceptron","provider_label":"perceptron","release_date":"2026-05-12","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"perceptron/perceptron-mk1-20260512","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":167,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"perceptron/perceptron-mk1.5-20260925","catalog_created_at":"2026-09-25T16:11:01+00:00","catalog_status":"active","context_length":36864,"eligible":false,"extra_pricing":{},"id":"perceptron/perceptron-mk1.5","input_price":0.15,"max_output_tokens":8192,"name":"Perceptron: Perceptron Mk1.5","new_model":false,"open_weights":false,"output_price":1.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"perceptron","provider_label":"perceptron","release_date":"2026-09-25","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"perceptron/perceptron-mk1.5-20260925","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":31,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"perplexity/sonar","catalog_created_at":"2025-01-27T21:36:48+00:00","catalog_status":"active","context_length":127072,"eligible":false,"extra_pricing":{},"id":"perplexity/sonar","input_price":1.0,"max_output_tokens":114364,"name":"Perplexity: Sonar","new_model":false,"open_weights":false,"output_price":1.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"perplexity","provider_label":"perplexity","release_date":"2025-01-27","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"perplexity/sonar","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":637,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"perplexity/sonar-deep-research","catalog_created_at":"2025-03-07T01:34:06+00:00","catalog_status":"active","context_length":128000,"eligible":false,"extra_pricing":{},"id":"perplexity/sonar-deep-research","input_price":2.0,"max_output_tokens":115200,"name":"Perplexity: Sonar Deep Research","new_model":false,"open_weights":false,"output_price":8.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"perplexity","provider_label":"perplexity","release_date":"2025-03-07","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"perplexity/sonar-deep-research","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":598,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"perplexity/sonar-pro","catalog_created_at":"2025-03-07T01:53:43+00:00","catalog_status":"active","context_length":200000,"eligible":false,"extra_pricing":{},"id":"perplexity/sonar-pro","input_price":3.0,"max_output_tokens":8000,"name":"Perplexity: Sonar Pro","new_model":false,"open_weights":false,"output_price":15.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"perplexity","provider_label":"perplexity","release_date":"2025-03-07","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"perplexity/sonar-pro","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":598,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"perplexity/sonar-pro-search","catalog_created_at":"2025-10-30T19:59:26+00:00","catalog_status":"active","context_length":200000,"eligible":false,"extra_pricing":{},"id":"perplexity/sonar-pro-search","input_price":3.0,"max_output_tokens":8000,"name":"Perplexity: Sonar Pro Search","new_model":false,"open_weights":false,"output_price":15.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"perplexity","provider_label":"perplexity","release_date":"2025-10-30","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"perplexity/sonar-pro-search","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":361,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"perplexity/sonar-reasoning-pro","catalog_created_at":"2025-03-07T02:08:28+00:00","catalog_status":"active","context_length":128000,"eligible":false,"extra_pricing":{},"id":"perplexity/sonar-reasoning-pro","input_price":2.0,"max_output_tokens":115200,"name":"Perplexity: Sonar Reasoning Pro","new_model":false,"open_weights":false,"output_price":8.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"perplexity","provider_label":"perplexity","release_date":"2025-03-07","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"perplexity/sonar-reasoning-pro","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":598,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.009,"canonical_slug":"poolside/laguna-s-2.1-20260720","catalog_created_at":"2026-07-21T16:51:23+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"poolside/laguna-s-2.1","input_price":0.09,"max_output_tokens":131072,"name":"Poolside: Laguna S 2.1","new_model":false,"open_weights":true,"output_price":0.18,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"poolside","provider_label":"poolside","release_date":"2026-07-21","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"poolside/laguna-s-2.1-20260720","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":97,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.03,"canonical_slug":"poolside/laguna-xs-2.1-20260625","catalog_created_at":"2026-07-02T14:27:09+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"poolside/laguna-xs-2.1","input_price":0.06,"max_output_tokens":32768,"name":"Poolside: Laguna XS 2.1","new_model":false,"open_weights":true,"output_price":0.12,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"poolside","provider_label":"poolside","release_date":"2026-07-02","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"poolside/laguna-xs-2.1-20260625","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":116,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.0375,"canonical_slug":"prism-ml/ternary-bonsai-2-27b-20260918","catalog_created_at":"2026-09-18T17:54:06+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"prism-ml/ternary-bonsai-2-27b","input_price":0.075,"max_output_tokens":32768,"name":"PrismML: Ternary Bonsai 2 27B","new_model":false,"open_weights":true,"output_price":0.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"prism-ml","provider_label":"prism-ml","release_date":"2026-09-18","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"prism-ml/ternary-bonsai-2-27b-20260918","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":38,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"qwen/qwen-2.5-72b-instruct","catalog_created_at":"2024-09-19T00:00:00+00:00","catalog_status":"active","context_length":32768,"eligible":false,"extra_pricing":{},"id":"qwen/qwen-2.5-72b-instruct","input_price":0.36,"max_output_tokens":16384,"name":"Qwen2.5 72B Instruct","new_model":false,"open_weights":true,"output_price":0.4,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2024-09-19","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen-2.5-72b-instruct","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":767,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"qwen/qwen-2.5-coder-32b-instruct","catalog_created_at":"2024-11-11T23:40:00+00:00","catalog_status":"active","context_length":32768,"eligible":false,"extra_pricing":{},"id":"qwen/qwen-2.5-coder-32b-instruct","input_price":0.66,"max_output_tokens":29491,"name":"Qwen2.5 Coder 32B Instruct","new_model":false,"open_weights":true,"output_price":1.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2024-11-12","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen-2.5-coder-32b-instruct","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":713,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"qwen/qwen-2.5-7b-instruct","catalog_created_at":"2024-10-16T00:00:00+00:00","catalog_status":"active","context_length":32768,"eligible":false,"extra_pricing":{},"id":"qwen/qwen-2.5-7b-instruct","input_price":0.1,"max_output_tokens":29491,"name":"Qwen: Qwen2.5 7B Instruct","new_model":false,"open_weights":true,"output_price":0.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2024-10-16","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen-2.5-7b-instruct","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":740,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.4,"canonical_slug":"qwen/qwen2.5-vl-72b-instruct","catalog_created_at":"2025-02-01T11:45:11+00:00","catalog_status":"active","context_length":128000,"eligible":false,"extra_pricing":{},"id":"qwen/qwen2.5-vl-72b-instruct","input_price":0.8,"max_output_tokens":115200,"name":"Qwen: Qwen2.5 VL 72B Instruct","new_model":false,"open_weights":true,"output_price":1.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2025-01-26","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen2.5-vl-72b-instruct","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":638,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"qwen/qwen3-14b-04-28","catalog_created_at":"2025-04-28T21:41:18+00:00","catalog_status":"active","context_length":40960,"eligible":false,"extra_pricing":{},"id":"qwen/qwen3-14b","input_price":0.12,"max_output_tokens":16384,"name":"Qwen: Qwen3 14B","new_model":false,"open_weights":true,"output_price":0.24,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2025-04-28","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3-14b-04-28","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":546,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"qwen/qwen3-coder-30b-a3b-instruct","catalog_created_at":"2025-07-31T14:32:59+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"qwen/qwen3-coder-30b-a3b-instruct","input_price":0.07,"max_output_tokens":235929,"name":"Qwen: Qwen3 Coder 30B A3B Instruct","new_model":false,"open_weights":true,"output_price":0.28,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2025-07-31","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3-coder-30b-a3b-instruct","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":452,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.1,"canonical_slug":"qwen/qwen3-coder-480b-a35b-07-25","catalog_created_at":"2025-07-23T00:29:06+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"qwen/qwen3-coder","input_price":0.3,"max_output_tokens":65536,"name":"Qwen: Qwen3 Coder 480B A35B","new_model":false,"open_weights":true,"output_price":1.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2025-07-22","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3-coder-480b-a35b-07-25","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":461,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.039,"canonical_slug":"qwen/qwen3-coder-flash","catalog_created_at":"2025-09-17T13:25:36+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"qwen/qwen3-coder-flash","input_price":0.195,"max_output_tokens":65536,"name":"Qwen: Qwen3 Coder Flash","new_model":false,"open_weights":false,"output_price":0.975,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2025-07-28","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3-coder-flash","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":455,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.07,"canonical_slug":"qwen/qwen3-coder-next-2025-02-03","catalog_created_at":"2026-02-04T00:15:01+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"qwen/qwen3-coder-next","input_price":0.12,"max_output_tokens":235929,"name":"Qwen: Qwen3 Coder Next","new_model":false,"open_weights":true,"output_price":0.8,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2026-02-03","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3-coder-next-2025-02-03","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":265,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"qwen/qwen3-vl-30b-a3b-instruct","catalog_created_at":"2025-10-06T23:47:56+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"qwen/qwen3-vl-30b-a3b-instruct","input_price":0.15,"max_output_tokens":16384,"name":"Qwen: Qwen3 VL 30B A3B Instruct","new_model":false,"open_weights":true,"output_price":0.6,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2025-10-04","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3-vl-30b-a3b-instruct","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":387,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"qwen/qwen3-vl-30b-a3b-thinking","catalog_created_at":"2025-10-06T23:47:59+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"qwen/qwen3-vl-30b-a3b-thinking","input_price":0.29,"max_output_tokens":235929,"name":"Qwen: Qwen3 VL 30B A3B Thinking","new_model":false,"open_weights":true,"output_price":1.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2025-10-06","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3-vl-30b-a3b-thinking","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":385,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.12,"canonical_slug":"qwen/qwen3-vl-8b-instruct","catalog_created_at":"2025-10-14T17:35:08+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"qwen/qwen3-vl-8b-instruct","input_price":0.25,"max_output_tokens":235929,"name":"Qwen: Qwen3 VL 8B Instruct","new_model":false,"open_weights":true,"output_price":0.75,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2025-10-14","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3-vl-8b-instruct","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":377,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"qwen/qwen3.5-plus-20260216","catalog_created_at":"2026-02-16T08:10:16+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"qwen/qwen3.5-plus-02-15","input_price":0.26,"max_output_tokens":65536,"name":"Qwen: Qwen3.5 Plus 2026-02-15","new_model":false,"open_weights":false,"output_price":1.56,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2026-02-16","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3.5-plus-20260216","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":252,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"qwen/qwen3.5-plus-20260420","catalog_created_at":"2026-04-27T03:42:48+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"qwen/qwen3.5-plus-20260420","input_price":0.3,"max_output_tokens":65536,"name":"Qwen: Qwen3.5 Plus 2026-04-20","new_model":false,"open_weights":false,"output_price":1.8,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2026-04-27","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3.5-plus-20260420","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":182,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"qwen/qwen3.5-9b-20260310","catalog_created_at":"2026-03-10T14:19:56+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"qwen/qwen3.5-9b","input_price":0.17,"max_output_tokens":235929,"name":"Qwen: Qwen3.5-9B","new_model":false,"open_weights":true,"output_price":0.25,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2026-02-23","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3.5-9b-20260310","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":245,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"qwen/qwen3.6-27b-20260422","catalog_created_at":"2026-04-27T01:57:44+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"qwen/qwen3.6-27b","input_price":0.32,"max_output_tokens":81920,"name":"Qwen: Qwen3.6 27B","new_model":false,"open_weights":true,"output_price":3.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2026-04-22","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3.6-27b-20260422","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":147.65,"ci_low":144.13,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-04-22"],"catalog_release_date":"2026-04-22","source_model_date":"2026-04-22","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"qwen/qwen3.6-27b-20260422","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":79,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":146.23,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-04-22","source_model_id":"Qwen3.6 27B","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":187,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.05,"canonical_slug":"qwen/qwen3.6-35b-a3b-20260415","catalog_created_at":"2026-04-27T03:24:15+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"qwen/qwen3.6-35b-a3b","input_price":0.1,"max_output_tokens":235929,"name":"Qwen: Qwen3.6 35B A3B","new_model":false,"open_weights":true,"output_price":0.95,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2026-04-17","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3.6-35b-a3b-20260415","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":192,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"qwen/qwen3.6-flash","catalog_created_at":"2026-04-27T03:42:42+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"qwen/qwen3.6-flash","input_price":0.1875,"max_output_tokens":65536,"name":"Qwen: Qwen3.6 Flash","new_model":false,"open_weights":false,"output_price":1.125,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2026-04-27","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3.6-flash","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":144.45,"ci_low":140.23,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":[],"catalog_release_date":"2026-04-27","source_model_date":"2026-04-27","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"qwen/qwen3.6-flash","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":104,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":142.89,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-04-27","source_model_id":"Qwen 3.6 Flash","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":182,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.006,"canonical_slug":"qwen/qwen3.7-flash-20260727","catalog_created_at":"2026-07-27T22:16:01+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"qwen/qwen3.7-flash","input_price":0.03,"max_output_tokens":65536,"name":"Qwen: Qwen3.7 Flash","new_model":false,"open_weights":false,"output_price":0.13,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2026-07-15","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3.7-flash-20260727","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":103,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.295,"canonical_slug":"qwen/qwen3.7-max-20260520","catalog_created_at":"2026-05-21T15:21:01+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"qwen/qwen3.7-max","input_price":1.475,"max_output_tokens":131072,"name":"Qwen: Qwen3.7 Max","new_model":false,"open_weights":false,"output_price":4.425,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2026-05-21","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3.7-max-20260520","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":158,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.25,"canonical_slug":"qwen/qwen3.8-2.4t-a95b-20260812","catalog_created_at":"2026-08-12T16:21:42+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"qwen/qwen3.8-2.4t-a95b","input_price":2.0,"max_output_tokens":909000,"name":"Qwen: Qwen3.8 2.4T A95B","new_model":false,"open_weights":true,"output_price":6.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2026-08-12","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3.8-2.4t-a95b-20260812","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":75,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.016,"canonical_slug":"qwen/qwen3.8-flash-20260826","catalog_created_at":"2026-08-26T19:37:40+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"qwen/qwen3.8-flash","input_price":0.15,"max_output_tokens":131072,"name":"Qwen: Qwen3.8 Flash","new_model":false,"open_weights":false,"output_price":0.47,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2026-08-26","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3.8-flash-20260826","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":61,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.5,"canonical_slug":"qwen/qwen3.8-max-prime-20260923","catalog_created_at":"2026-09-23T19:20:28+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"qwen/qwen3.8-max-prime","input_price":4.0,"max_output_tokens":131072,"name":"Qwen: Qwen3.8 Max Prime","new_model":false,"open_weights":false,"output_price":12.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2026-09-23","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3.8-max-prime-20260923","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":33,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.016,"canonical_slug":"qwen/qwen3.8-omni-flash-20260918","catalog_created_at":"2026-09-21T03:05:54+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"qwen/qwen3.8-omni-flash","input_price":0.15,"max_output_tokens":131072,"name":"Qwen: Qwen3.8 Omni Flash","new_model":false,"open_weights":false,"output_price":0.47,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":"2026-09-17","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3.8-omni-flash-20260918","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":39,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"rekaai/reka-edge-2603","catalog_created_at":"2026-03-20T17:16:05+00:00","catalog_status":"active","context_length":16384,"eligible":false,"extra_pricing":{},"id":"rekaai/reka-edge","input_price":0.1,"max_output_tokens":14745,"name":"Reka Edge","new_model":false,"open_weights":true,"output_price":0.1,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"rekaai","provider_label":"rekaai","release_date":"2026-03-11","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"rekaai/reka-edge-2603","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":229,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"rekaai/reka-flash-3","catalog_created_at":"2025-03-12T20:53:33+00:00","catalog_status":"active","context_length":65536,"eligible":false,"extra_pricing":{},"id":"rekaai/reka-flash-3","input_price":0.1,"max_output_tokens":58982,"name":"Reka Flash 3","new_model":false,"open_weights":true,"output_price":0.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"rekaai","provider_label":"rekaai","release_date":"2025-03-10","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"rekaai/reka-flash-3","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":595,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"relace/relace-apply-3","catalog_created_at":"2025-09-26T12:59:32+00:00","catalog_status":"active","context_length":256000,"eligible":false,"extra_pricing":{},"id":"relace/relace-apply-3","input_price":0.85,"max_output_tokens":128000,"name":"Relace: Relace Apply 3","new_model":false,"open_weights":false,"output_price":1.25,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"relace","provider_label":"relace","release_date":"2025-09-26","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"relace/relace-apply-3","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":395,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"relace/relace-search-20251208","catalog_created_at":"2025-12-08T17:06:00+00:00","catalog_status":"active","context_length":256000,"eligible":false,"extra_pricing":{},"id":"relace/relace-search","input_price":1.0,"max_output_tokens":128000,"name":"Relace: Relace Search","new_model":false,"open_weights":false,"output_price":3.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"relace","provider_label":"relace","release_date":"2025-12-08","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"relace/relace-search-20251208","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":322,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.25,"canonical_slug":"sakana/fugu-max-20260911","catalog_created_at":"2026-09-11T05:32:51+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"sakana/fugu-max","input_price":2.0,"max_output_tokens":128000,"name":"Sakana: Fugu Max","new_model":false,"open_weights":false,"output_price":6.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"sakana","provider_label":"sakana","release_date":"2026-09-11","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"sakana/fugu-max-20260911","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":45,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.5,"canonical_slug":"sakana/fugu-ultra-20260615","catalog_created_at":"2026-06-24T04:45:03+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"sakana/fugu-ultra","input_price":5.0,"max_output_tokens":128000,"name":"Sakana: Fugu Ultra","new_model":false,"open_weights":false,"output_price":30.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"sakana","provider_label":"sakana","release_date":"2026-06-15","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"sakana/fugu-ultra-20260615","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":133,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.5,"canonical_slug":"sakana/fugu-ultra-v2-20260911","catalog_created_at":"2026-09-11T05:43:03+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"sakana/fugu-ultra-v2","input_price":5.0,"max_output_tokens":128000,"name":"Sakana: Fugu Ultra v2","new_model":false,"open_weights":false,"output_price":30.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"sakana","provider_label":"sakana","release_date":"2026-09-11","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"sakana/fugu-ultra-v2-20260911","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":45,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.15,"canonical_slug":"sakana/namazu-20260811","catalog_created_at":"2026-08-11T01:02:09+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"sakana/sakana-namazu","input_price":0.95,"max_output_tokens":65536,"name":"Sakana: Sakana Namazu","new_model":false,"open_weights":false,"output_price":4.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"sakana","provider_label":"sakana","release_date":"2026-08-03","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"sakana/namazu-20260811","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":84,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"sao10k/l3-lunaris-8b","catalog_created_at":"2024-08-13T00:00:00+00:00","catalog_status":"active","context_length":8192,"eligible":false,"extra_pricing":{},"id":"sao10k/l3-lunaris-8b","input_price":0.04,"max_output_tokens":7372,"name":"Sao10K: Llama 3 8B Lunaris","new_model":false,"open_weights":true,"output_price":0.05,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"sao10k","provider_label":"sao10k","release_date":"2024-08-13","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"sao10k/l3-lunaris-8b","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":804,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"sao10k/l3.1-euryale-70b","catalog_created_at":"2024-08-28T00:00:00+00:00","catalog_status":"active","context_length":131072,"eligible":false,"extra_pricing":{},"id":"sao10k/l3.1-euryale-70b","input_price":0.85,"max_output_tokens":16384,"name":"Sao10K: Llama 3.1 Euryale 70B v2.2","new_model":false,"open_weights":true,"output_price":0.85,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"sao10k","provider_label":"sao10k","release_date":"2024-08-28","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"sao10k/l3.1-euryale-70b","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":789,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.04,"canonical_slug":"stepfun/step-3.7-flash-20260528","catalog_created_at":"2026-05-28T16:17:49+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"stepfun/step-3.7-flash","input_price":0.2,"max_output_tokens":230400,"name":"StepFun: Step 3.7 Flash","new_model":false,"open_weights":true,"output_price":1.15,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"stepfun","provider_label":"stepfun","release_date":"2026-05-29","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"stepfun/step-3.7-flash-20260528","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":150,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"tencent/hunyuan-a13b-instruct","catalog_created_at":"2025-07-08T15:14:24+00:00","catalog_status":"active","context_length":131072,"eligible":false,"extra_pricing":{},"id":"tencent/hunyuan-a13b-instruct","input_price":0.14,"max_output_tokens":117964,"name":"Tencent: Hunyuan A13B Instruct","new_model":false,"open_weights":true,"output_price":0.57,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"tencent","provider_label":"tencent","release_date":"2025-07-08","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"tencent/hunyuan-a13b-instruct","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":475,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"tencent/hy-mt2-1.8b-20260521","catalog_created_at":"2026-08-20T13:13:01+00:00","catalog_status":"active","context_length":8192,"eligible":false,"extra_pricing":{},"id":"tencent/hy-mt2-1.8b","input_price":0.044,"max_output_tokens":4096,"name":"Tencent: Hy-MT2-1.8B","new_model":false,"open_weights":true,"output_price":0.177,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"tencent","provider_label":"tencent","release_date":"2026-05-21","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"tencent/hy-mt2-1.8b-20260521","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":158,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"tencent/hy-mt2-30b-a3b-20260521","catalog_created_at":"2026-08-20T13:12:41+00:00","catalog_status":"active","context_length":8192,"eligible":false,"extra_pricing":{},"id":"tencent/hy-mt2-30b-a3b","input_price":0.074,"max_output_tokens":4096,"name":"Tencent: Hy-MT2-30B-A3B","new_model":false,"open_weights":true,"output_price":0.295,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"tencent","provider_label":"tencent","release_date":"2026-05-21","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"tencent/hy-mt2-30b-a3b-20260521","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":158,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"tencent/hy-mt2-7b-20260521","catalog_created_at":"2026-08-19T14:13:17+00:00","catalog_status":"active","context_length":8192,"eligible":false,"extra_pricing":{},"id":"tencent/hy-mt2-7b","input_price":0.074,"max_output_tokens":4096,"name":"Tencent: Hy-MT2-7B","new_model":false,"open_weights":true,"output_price":0.295,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"tencent","provider_label":"tencent","release_date":"2026-05-21","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"tencent/hy-mt2-7b-20260521","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":158,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.0378,"canonical_slug":"tencent/hy4-preview-20260827","catalog_created_at":"2026-08-28T06:09:35+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"tencent/hy4-preview","input_price":0.7506,"max_output_tokens":64000,"name":"Tencent: Hy4 preview","new_model":false,"open_weights":true,"output_price":2.2509,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"tencent","provider_label":"tencent","release_date":"2026-08-28","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"tencent/hy4-preview-20260827","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[{"id":"webdev","label":"做网页","date":"2026-10-23"}],"effort":null,"mapping_type":null,"listed_days_ago":59,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.15,"canonical_slug":"thedrummer/cydonia-24b-v4.1","catalog_created_at":"2025-09-27T00:11:18+00:00","catalog_status":"active","context_length":131072,"eligible":false,"extra_pricing":{},"id":"thedrummer/cydonia-24b-v4.1","input_price":0.3,"max_output_tokens":117964,"name":"TheDrummer: Cydonia 24B V4.1","new_model":false,"open_weights":true,"output_price":0.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"thedrummer","provider_label":"thedrummer","release_date":"2025-09-27","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"thedrummer/cydonia-24b-v4.1","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":394,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.25,"canonical_slug":"thedrummer/skyfall-36b-v2","catalog_created_at":"2025-03-10T19:56:06+00:00","catalog_status":"active","context_length":32768,"eligible":false,"extra_pricing":{},"id":"thedrummer/skyfall-36b-v2","input_price":0.55,"max_output_tokens":29491,"name":"TheDrummer: Skyfall 36B V2","new_model":false,"open_weights":true,"output_price":0.8,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"thedrummer","provider_label":"thedrummer","release_date":"2025-03-10","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"thedrummer/skyfall-36b-v2","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":595,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"thedrummer/unslopnemo-12b","catalog_created_at":"2024-11-08T22:04:08+00:00","catalog_status":"active","context_length":1024000,"eligible":false,"extra_pricing":{},"id":"thedrummer/unslopnemo-12b","input_price":0.4,"max_output_tokens":819200,"name":"TheDrummer: UnslopNemo 12B","new_model":false,"open_weights":true,"output_price":0.4,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"thedrummer","provider_label":"thedrummer","release_date":"2024-11-08","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"thedrummer/unslopnemo-12b","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":717,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"typesafe/jev-router","catalog_created_at":"2026-09-25T19:12:40+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"typesafe/jev-router","input_price":null,"max_output_tokens":65536,"name":"TypeSafe: Jev Router","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"typesafe","provider_label":"typesafe","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"typesafe/jev-router","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":31,"new_model_basis":"catalog_created_at","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"undi95/remm-slerp-l2-13b","catalog_created_at":"2023-07-22T00:00:00+00:00","catalog_status":"active","context_length":6144,"eligible":false,"extra_pricing":{},"id":"undi95/remm-slerp-l2-13b","input_price":0.35,"max_output_tokens":5529,"name":"ReMM SLERP 13B","new_model":false,"open_weights":true,"output_price":0.65,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"undi95","provider_label":"undi95","release_date":"2023-07-22","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"undi95/remm-slerp-l2-13b","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":1192,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.005,"canonical_slug":"upstage/solar-mini4-20260922","catalog_created_at":"2026-09-23T10:45:58+00:00","catalog_status":"active","context_length":524288,"eligible":false,"extra_pricing":{},"id":"upstage/solar-mini4","input_price":0.05,"max_output_tokens":131072,"name":"Upstage: Solar Mini 4","new_model":false,"open_weights":false,"output_price":0.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"upstage","provider_label":"upstage","release_date":"2026-09-22","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"upstage/solar-mini4-20260922","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":34,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.015,"canonical_slug":"upstage/solar-pro-3","catalog_created_at":"2026-01-27T02:33:20+00:00","catalog_status":"active","context_length":131072,"eligible":false,"extra_pricing":{},"id":"upstage/solar-pro-3","input_price":0.15,"max_output_tokens":117964,"name":"Upstage: Solar Pro 3","new_model":false,"open_weights":false,"output_price":0.6,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"upstage","provider_label":"upstage","release_date":"2026-01-26","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"upstage/solar-pro-3","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":273,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.03,"canonical_slug":"upstage/solar-pro4-20260810","catalog_created_at":"2026-08-10T14:20:36+00:00","catalog_status":"active","context_length":524288,"eligible":false,"extra_pricing":{},"id":"upstage/solar-pro4","input_price":0.15,"max_output_tokens":131072,"name":"Upstage: Solar Pro 4","new_model":false,"open_weights":false,"output_price":0.6,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"upstage","provider_label":"upstage","release_date":"2026-08-06","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"upstage/solar-pro4-20260810","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":81,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"writer/palmyra-x5-20250428","catalog_created_at":"2026-01-21T13:57:03+00:00","catalog_status":"active","context_length":1040000,"eligible":false,"extra_pricing":{},"id":"writer/palmyra-x5","input_price":0.6,"max_output_tokens":8192,"name":"Writer: Palmyra X5","new_model":false,"open_weights":false,"output_price":6.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"writer","provider_label":"writer","release_date":"2025-04-28","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"writer/palmyra-x5-20250428","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":546,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.2,"canonical_slug":"x-ai/grok-4.20-20260309","catalog_created_at":"2026-03-31T17:43:39+00:00","catalog_status":"active","context_length":2000000,"eligible":false,"extra_pricing":{},"id":"x-ai/grok-4.20","input_price":1.25,"max_output_tokens":1800000,"name":"SpaceXAI: Grok 4.20","new_model":false,"open_weights":false,"output_price":2.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"x-ai","provider_label":"x-ai","release_date":"2026-03-31","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"x-ai/grok-4.20-20260309","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":209,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.2,"canonical_slug":"x-ai/grok-4.20-multi-agent-20260309","catalog_created_at":"2026-03-31T17:45:58+00:00","catalog_status":"active","context_length":2000000,"eligible":false,"extra_pricing":{},"id":"x-ai/grok-4.20-multi-agent","input_price":1.25,"max_output_tokens":1800000,"name":"SpaceXAI: Grok 4.20 Multi-Agent","new_model":false,"open_weights":false,"output_price":2.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"x-ai","provider_label":"x-ai","release_date":"2026-03-31","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"x-ai/grok-4.20-multi-agent-20260309","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":209,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.2,"canonical_slug":"x-ai/grok-build-0.1-20260520","catalog_created_at":"2026-05-20T17:28:43+00:00","catalog_status":"active","context_length":256000,"eligible":false,"extra_pricing":{},"id":"x-ai/grok-build-0.1","input_price":1.0,"max_output_tokens":230400,"name":"SpaceXAI: Grok Build 0.1","new_model":false,"open_weights":false,"output_price":2.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"x-ai","provider_label":"x-ai","release_date":"2026-04-16","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"x-ai/grok-build-0.1-20260520","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":193,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.036,"canonical_slug":"xiaomi/mimo-v2.6-pro-ultraspeed-20260921","catalog_created_at":"2026-09-21T20:07:47+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"xiaomi/mimo-v2.6-pro-ultraspeed","input_price":4.35,"max_output_tokens":131072,"name":"Xiaomi: MiMo-V2.6-Pro-UltraSpeed","new_model":false,"open_weights":true,"output_price":8.7,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"xiaomi","provider_label":"xiaomi","release_date":"2026-09-21","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"xiaomi/mimo-v2.6-pro-ultraspeed-20260921","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":35,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.24,"canonical_slug":"z-ai/glm-5-turbo-20260315","catalog_created_at":"2026-03-15T14:06:13+00:00","catalog_status":"active","context_length":202752,"eligible":false,"extra_pricing":{},"id":"z-ai/glm-5-turbo","input_price":1.2,"max_output_tokens":131072,"name":"Z.ai: GLM 5 Turbo","new_model":false,"open_weights":false,"output_price":4.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"z-ai","provider_label":"z-ai","release_date":"2026-03-16","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"z-ai/glm-5-turbo-20260315","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":224,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.09,"canonical_slug":"z-ai/glm-5.3-flashx-20260918","catalog_created_at":"2026-09-18T15:07:00+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"z-ai/glm-5.3-flashx","input_price":0.37,"max_output_tokens":131072,"name":"Z.ai: GLM 5.3 FlashX","new_model":false,"open_weights":false,"output_price":1.25,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"z-ai","provider_label":"z-ai","release_date":"2026-09-18","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"z-ai/glm-5.3-flashx-20260918","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":38,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.56,"canonical_slug":"z-ai/glm-5.3-prime-20260921","catalog_created_at":"2026-09-23T21:40:51+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"z-ai/glm-5.3-prime","input_price":2.8,"max_output_tokens":131072,"name":"Z.ai: GLM 5.3 Prime","new_model":false,"open_weights":false,"output_price":8.8,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"z-ai","provider_label":"z-ai","release_date":"2026-09-23","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"z-ai/glm-5.3-prime-20260921","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":33,"new_model_basis":"release_date","variant":false},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"anthropic/claude-haiku-5.5-20261007","catalog_created_at":"2026-10-07T18:31:23+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"anthropic/claude-haiku-5.5:batch","input_price":null,"max_output_tokens":128000,"name":"Anthropic: Claude Haiku 5.5 (batch)","new_model":true,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthropic","provider_label":"anthropic","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthropic/claude-haiku-5.5-20261007","version_policy":"independent_snapshot","evaluation_references":[{"benchmark":"Vals Index","benchmark_version":"2.2","checked_at":"2026-10-26T00:20:45.595508+00:00","date":"2026-10-24","effort":null,"identity_evidence":{"checked_at":"2026-10-26T00:20:45.595508+00:00","effort":null,"release_date":"2026-10-07","sha256":"06ea7e0887d58d20f5080ecbbb8b973f0471a6ee2eee78814bc7b478b1f97ad6","source_name":"Claude Haiku 5.5","url":"https://www.vals.ai/models/anthropic_claude-haiku-5-5"},"kind":"independent","model_url":"https://www.vals.ai/models/anthropic_claude-haiku-5-5","model_version":"anthropic/claude-haiku-5.5-20261007","note":"按经济领域权重汇总的任务成绩；原榜含部分其他模型使用备用模型的结果。推理档位为来源模型页默认值，单项测试可能不同，完整条件见原榜。","rank":12,"scope":"该版本综合榜全部已评测模型","score":58.71,"source_id":"vals","source_model_id":"Claude Haiku 5.5","source_url":"https://www.vals.ai/benchmarks/vals_index","total":27,"unit":"%","update_mode":"automatic","stale":false},{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":162.47,"ci_low":156.19,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-10-07"],"catalog_release_date":"2026-10-07","source_model_date":"2026-10-07","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"anthropic/claude-haiku-5.5-20261007","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":16,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":159.32,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-10-07","source_model_id":"Claude Haiku 5.5","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":19,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"anthropic/claude-sonnet-5.5-20260928","catalog_created_at":"2026-09-28T18:04:46+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"anthropic/claude-sonnet-5.5:batch","input_price":null,"max_output_tokens":128000,"name":"Anthropic: Claude Sonnet 5.5 (batch)","new_model":true,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthropic","provider_label":"anthropic","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthropic/claude-sonnet-5.5-20260928","version_policy":"independent_snapshot","evaluation_references":[{"benchmark":"Vals Index","benchmark_version":"2.2","checked_at":"2026-10-26T00:20:45.595508+00:00","date":"2026-10-24","effort":null,"identity_evidence":{"checked_at":"2026-10-26T00:20:45.595508+00:00","effort":null,"release_date":"2026-09-28","sha256":"d3f5e9342259a56564d40f8b2c059b4e3865d7d214b8c0523c2ccb83b409748d","source_name":"Claude Sonnet 5.5","url":"https://www.vals.ai/models/anthropic_claude-sonnet-5-5"},"kind":"independent","model_url":"https://www.vals.ai/models/anthropic_claude-sonnet-5-5","model_version":"anthropic/claude-sonnet-5.5-20260928","note":"按经济领域权重汇总的任务成绩；原榜含部分其他模型使用备用模型的结果。推理档位为来源模型页默认值，单项测试可能不同，完整条件见原榜。","rank":1,"scope":"该版本综合榜全部已评测模型","score":70.85,"source_id":"vals","source_model_id":"Claude Sonnet 5.5","source_url":"https://www.vals.ai/benchmarks/vals_index","total":27,"unit":"%","update_mode":"automatic","stale":false},{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":171.34,"ci_low":163.06,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-09-28"],"catalog_release_date":"2026-09-28","source_model_date":"2026-09-28","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"anthropic/claude-sonnet-5.5-20260928","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":6,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":166.18,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-09-28","source_model_id":"Claude Sonnet 5.5","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":28,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"google/gemini-4-argon-20261013","catalog_created_at":"2026-10-13T16:04:37+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"google/gemini-4-argon:batch","input_price":null,"max_output_tokens":943718,"name":"Google: Gemini 4 Argon (batch)","new_model":true,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-4-argon-20261013","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":13,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"nvidia/nemotron-3.5-super-vl-20261015","catalog_created_at":"2026-10-15T15:04:22+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"nvidia/nemotron-3.5-super-vl:free","input_price":0.0,"max_output_tokens":131072,"name":"NVIDIA: Nemotron 3.5 Super VL (free)","new_model":true,"open_weights":true,"output_price":0.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"nvidia","provider_label":"nvidia","release_date":"2026-10-15","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"nvidia/nemotron-3.5-super-vl-20261015","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":11,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-6.1-sol-20260929","catalog_created_at":"2026-09-29T17:28:02+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-6.1-sol:batch","input_price":null,"max_output_tokens":128000,"name":"OpenAI: GPT-6.1 Sol (batch)","new_model":true,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-6.1-sol-20260929","version_policy":"independent_snapshot","evaluation_references":[{"benchmark":"Vals Index","benchmark_version":"2.2","checked_at":"2026-10-26T00:20:45.595508+00:00","date":"2026-10-24","effort":"max","identity_evidence":{"checked_at":"2026-10-26T00:20:45.595508+00:00","effort":"max","release_date":"2026-09-29","sha256":"821f4fc589e98e0eab9464c518b6bc28b954f5fef29665ed26d19b5b4a1ea7be","source_name":"GPT-6.1 Sol","url":"https://www.vals.ai/models/openai_gpt-6.1-sol"},"kind":"independent","model_url":"https://www.vals.ai/models/openai_gpt-6.1-sol","model_version":"openai/gpt-6.1-sol-20260929","note":"按经济领域权重汇总的任务成绩；原榜含部分其他模型使用备用模型的结果。推理档位为来源模型页默认值，单项测试可能不同，完整条件见原榜。","rank":7,"scope":"该版本综合榜全部已评测模型","score":63.94,"source_id":"vals","source_model_id":"GPT-6.1 Sol","source_url":"https://www.vals.ai/benchmarks/vals_index","total":27,"unit":"%","update_mode":"automatic","stale":false},{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":174.43,"ci_low":165.0,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-09-29"],"catalog_release_date":"2026-09-29","source_model_date":"2026-09-29","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-6.1-sol-20260929","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":3,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":168.37,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-09-29","source_model_id":"GPT-6.1 Sol","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":27,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-6.1-sol-pro-20260929","catalog_created_at":"2026-09-29T17:28:06+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-6.1-sol-pro:batch","input_price":null,"max_output_tokens":128000,"name":"OpenAI: GPT-6.1 Sol Pro (batch)","new_model":true,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-6.1-sol-pro-20260929","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":27,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"recursive/flint-v1.0-20261021","catalog_created_at":"2026-10-21T22:46:05+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"recursive/flint-v1.0:free","input_price":0.0,"max_output_tokens":235929,"name":"Recursive: Flint (free)","new_model":true,"open_weights":true,"output_price":0.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"recursive","provider_label":"recursive","release_date":"2026-10-21","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"recursive/flint-v1.0-20261021","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":5,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"xingchen-agi/xing4.0-29b-a4b-20261015","catalog_created_at":"2026-10-16T19:04:56+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"xingchen-agi/xing4.0-29b-a4b:free","input_price":0.0,"max_output_tokens":235929,"name":"XingChen: Xing4.0 29B A4B (free)","new_model":true,"open_weights":true,"output_price":0.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"xingchen-agi","provider_label":"xingchen-agi","release_date":"2026-10-16","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"xingchen-agi/xing4.0-29b-a4b-20261015","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":10,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.03,"canonical_slug":"~unbiased/pareto-latest","catalog_created_at":"2026-10-13T23:10:50+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"~unbiased/pareto-latest","input_price":0.8,"max_output_tokens":131072,"name":"Pareto Latest","new_model":true,"open_weights":false,"output_price":3.2,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"~unbiased","provider_label":"~unbiased","release_date":"2026-10-13","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"~unbiased/pareto-latest","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":13,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"anthropic/claude-5-fable-20260609","catalog_created_at":"2026-06-09T12:18:35+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"anthropic/claude-fable-5:batch","input_price":null,"max_output_tokens":128000,"name":"Anthropic: Claude Fable 5 (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthropic","provider_label":"anthropic","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthropic/claude-5-fable-20260609","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":168.18,"ci_low":160.53,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-06-09"],"catalog_release_date":"2026-06-09","source_model_date":"2026-06-09","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"anthropic/claude-5-fable-20260609","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":9,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":163.42,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-06-09","source_model_id":"Claude Fable 5","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":139,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"anthropic/claude-fable-5.1-20260831","catalog_created_at":"2026-09-01T18:03:58+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"anthropic/claude-fable-5.1:batch","input_price":null,"max_output_tokens":128000,"name":"Anthropic: Claude Fable 5.1 (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthropic","provider_label":"anthropic","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthropic/claude-fable-5.1-20260831","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":55,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"anthropic/claude-4.5-haiku-20251001","catalog_created_at":"2025-10-15T17:00:38+00:00","catalog_status":"active","context_length":200000,"eligible":false,"extra_pricing":{},"id":"anthropic/claude-haiku-4.5:batch","input_price":null,"max_output_tokens":64000,"name":"Anthropic: Claude Haiku 4.5 (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthropic","provider_label":"anthropic","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthropic/claude-4.5-haiku-20251001","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":376,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"anthropic/claude-4.1-opus-20250805","catalog_created_at":"2025-08-05T16:33:11+00:00","catalog_status":"active","context_length":200000,"eligible":false,"extra_pricing":{},"id":"anthropic/claude-opus-4.1:batch","input_price":null,"max_output_tokens":32000,"name":"Anthropic: Claude Opus 4.1 (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthropic","provider_label":"anthropic","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthropic/claude-4.1-opus-20250805","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":146.55,"ci_low":141.75,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-08-05"],"catalog_release_date":"2025-08-05","source_model_date":"2025-08-05","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"anthropic/claude-4.1-opus-20250805","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":95,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":144.07,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-08-05","source_model_id":"Claude Opus 4.1","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":447,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"anthropic/claude-4.5-opus-20251124","catalog_created_at":"2025-11-24T18:56:20+00:00","catalog_status":"active","context_length":200000,"eligible":false,"extra_pricing":{},"id":"anthropic/claude-opus-4.5:batch","input_price":null,"max_output_tokens":64000,"name":"Anthropic: Claude Opus 4.5 (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthropic","provider_label":"anthropic","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthropic/claude-4.5-opus-20251124","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":153.01,"ci_low":147.81,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-11-24"],"catalog_release_date":"2025-11-24","source_model_date":"2025-11-24","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"anthropic/claude-4.5-opus-20251124","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":58,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":150.11,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-11-24","source_model_id":"Claude Opus 4.5","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":336,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"anthropic/claude-4.6-opus-20260205","catalog_created_at":"2026-02-04T15:30:50+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"anthropic/claude-opus-4.6:batch","input_price":null,"max_output_tokens":128000,"name":"Anthropic: Claude Opus 4.6 (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthropic","provider_label":"anthropic","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthropic/claude-4.6-opus-20260205","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":158.07,"ci_low":153.82,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-02-05"],"catalog_release_date":"2026-02-05","source_model_date":"2026-02-05","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"anthropic/claude-4.6-opus-20260205","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":33,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":155.66,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-02-05","source_model_id":"Claude Opus 4.6","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":264,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"anthropic/claude-4.7-opus-20260416","catalog_created_at":"2026-04-16T14:51:40+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"anthropic/claude-opus-4.7:batch","input_price":null,"max_output_tokens":128000,"name":"Anthropic: Claude Opus 4.7 (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthropic","provider_label":"anthropic","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthropic/claude-4.7-opus-20260416","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":159.19,"ci_low":154.91,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-04-16"],"catalog_release_date":"2026-04-16","source_model_date":"2026-04-16","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"anthropic/claude-4.7-opus-20260416","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":26,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":156.79,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-04-16","source_model_id":"Claude Opus 4.7","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":193,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"anthropic/claude-4.8-opus-20260528","catalog_created_at":"2026-05-27T18:04:51+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"anthropic/claude-opus-4.8:batch","input_price":null,"max_output_tokens":128000,"name":"Anthropic: Claude Opus 4.8 (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthropic","provider_label":"anthropic","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthropic/claude-4.8-opus-20260528","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":161.95,"ci_low":156.89,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-05-28"],"catalog_release_date":"2026-05-28","source_model_date":"2026-05-28","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"anthropic/claude-4.8-opus-20260528","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":17,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":159.01,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-05-28","source_model_id":"Claude Opus 4.8","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":152,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"anthropic/claude-opus-5-20260723","catalog_created_at":"2026-07-24T17:02:24+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"anthropic/claude-opus-5:batch","input_price":null,"max_output_tokens":128000,"name":"Anthropic: Claude Opus 5 (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthropic","provider_label":"anthropic","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthropic/claude-opus-5-20260723","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":94,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"anthropic/claude-opus-5.5-20260921","catalog_created_at":"2026-09-22T16:32:12+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"anthropic/claude-opus-5.5:batch","input_price":null,"max_output_tokens":128000,"name":"Anthropic: Claude Opus 5.5 (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthropic","provider_label":"anthropic","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthropic/claude-opus-5.5-20260921","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":34,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"anthropic/claude-4.5-sonnet-20250929","catalog_created_at":"2025-09-29T16:01:16+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"anthropic/claude-sonnet-4.5:batch","input_price":null,"max_output_tokens":64000,"name":"Anthropic: Claude Sonnet 4.5 (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthropic","provider_label":"anthropic","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthropic/claude-4.5-sonnet-20250929","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":148.31,"ci_low":145.03,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-09-29"],"catalog_release_date":"2025-09-29","source_model_date":"2025-09-29","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"anthropic/claude-4.5-sonnet-20250929","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":75,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":146.74,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-09-29","source_model_id":"Claude Sonnet 4.5","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":392,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"anthropic/claude-4.6-sonnet-20260217","catalog_created_at":"2026-02-17T15:43:10+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"anthropic/claude-sonnet-4.6:batch","input_price":null,"max_output_tokens":128000,"name":"Anthropic: Claude Sonnet 4.6 (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthropic","provider_label":"anthropic","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthropic/claude-4.6-sonnet-20260217","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":155.06,"ci_low":150.2,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-02-17"],"catalog_release_date":"2026-02-17","source_model_date":"2026-02-17","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"anthropic/claude-4.6-sonnet-20260217","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":49,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":152.79,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-02-17","source_model_id":"Claude Sonnet 4.6","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":251,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"anthropic/claude-sonnet-5-20260630","catalog_created_at":"2026-06-30T18:11:23+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"anthropic/claude-sonnet-5:batch","input_price":null,"max_output_tokens":128000,"name":"Anthropic: Claude Sonnet 5 (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"anthropic","provider_label":"anthropic","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"anthropic/claude-sonnet-5-20260630","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":159.65,"ci_low":154.1,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-06-30"],"catalog_release_date":"2026-06-30","source_model_date":"2026-06-30","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"anthropic/claude-sonnet-5-20260630","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":25,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":156.88,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-06-30","source_model_id":"Claude Sonnet 5","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":118,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"cohere/north-mini-code-20260617","catalog_created_at":"2026-06-17T19:15:48+00:00","catalog_status":"active","context_length":256000,"eligible":false,"extra_pricing":{},"id":"cohere/north-mini-code:free","input_price":0.0,"max_output_tokens":64000,"name":"Cohere: North Mini Code (free)","new_model":false,"open_weights":true,"output_price":0.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"cohere","provider_label":"cohere","release_date":"2026-06-09","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"cohere/north-mini-code-20260617","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":139,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"deepseek/deepseek-v4.1-flash-20260910","catalog_created_at":"2026-09-10T06:21:25+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"deepseek/deepseek-v4.1-flash:batch","input_price":null,"max_output_tokens":384000,"name":"DeepSeek: DeepSeek V4.1 Flash (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"deepseek","provider_label":"deepseek","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"deepseek/deepseek-v4.1-flash-20260910","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":46,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"dots-studio/dots-3-note-preview-20260813","catalog_created_at":"2026-08-14T04:06:01+00:00","catalog_status":"active","context_length":512000,"eligible":false,"extra_pricing":{},"id":"dots-studio/dots-3-note-preview:free","input_price":0.0,"max_output_tokens":460800,"name":"Dots Studio: Dots3-Note Preview (free)","new_model":false,"open_weights":false,"output_price":0.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"dots-studio","provider_label":"dots-studio","release_date":"2026-08-14","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"dots-studio/dots-3-note-preview-20260813","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":73,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"google/gemini-3-flash-preview-20251217","catalog_created_at":"2025-12-17T15:57:58+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"google/gemini-3-flash-preview:batch","input_price":null,"max_output_tokens":65536,"name":"Google: Gemini 3 Flash Preview (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-3-flash-preview-20251217","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":313,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"google/gemini-3.1-flash-lite-20260507","catalog_created_at":"2026-05-07T15:47:08+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"google/gemini-3.1-flash-lite:batch","input_price":null,"max_output_tokens":65536,"name":"Google: Gemini 3.1 Flash Lite (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-3.1-flash-lite-20260507","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":172,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"google/gemini-3.1-pro-preview-20260219","catalog_created_at":"2026-02-19T14:00:27+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"google/gemini-3.1-pro-preview:batch","input_price":null,"max_output_tokens":65536,"name":"Google: Gemini 3.1 Pro Preview (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-3.1-pro-preview-20260219","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":249,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"google/gemini-3.5-flash-20260519","catalog_created_at":"2026-05-19T12:30:00+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"google/gemini-3.5-flash:batch","input_price":null,"max_output_tokens":65536,"name":"Google: Gemini 3.5 Flash (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-3.5-flash-20260519","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":157.41,"ci_low":152.91,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-05-19"],"catalog_release_date":"2026-05-19","source_model_date":"2026-05-19","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"google/gemini-3.5-flash-20260519","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":39,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":154.88,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-05-19","source_model_id":"Gemini 3.5 Flash","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":160,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"google/gemini-3.5-flash-lite-20260721","catalog_created_at":"2026-07-21T15:12:06+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"google/gemini-3.5-flash-lite:batch","input_price":null,"max_output_tokens":65536,"name":"Google: Gemini 3.5 Flash Lite (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-3.5-flash-lite-20260721","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":146.26,"ci_low":142.41,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-07-21"],"catalog_release_date":"2026-07-21","source_model_date":"2026-07-21","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"google/gemini-3.5-flash-lite-20260721","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":91,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":144.91,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-07-21","source_model_id":"Gemini 3.5 Flash-Lite","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":97,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"google/gemini-3.6-flash-20260721","catalog_created_at":"2026-07-21T15:12:13+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"google/gemini-3.6-flash:batch","input_price":null,"max_output_tokens":65536,"name":"Google: Gemini 3.6 Flash (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-3.6-flash-20260721","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":156.58,"ci_low":152.85,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-07-21"],"catalog_release_date":"2026-07-21","source_model_date":"2026-07-21","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"google/gemini-3.6-flash-20260721","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":41,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":154.53,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-07-21","source_model_id":"Gemini 3.6 Flash","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":97,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"google/gemini-3.7-flash-20260813","catalog_created_at":"2026-08-13T17:03:01+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"google/gemini-3.7-flash:batch","input_price":null,"max_output_tokens":65536,"name":"Google: Gemini 3.7 Flash (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-3.7-flash-20260813","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":161.24,"ci_low":155.92,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-08-13"],"catalog_release_date":"2026-08-13","source_model_date":"2026-08-13","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"google/gemini-3.7-flash-20260813","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":18,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":157.97,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-08-13","source_model_id":"Gemini 3.7 Flash","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":74,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"google/gemini-3.8-flash-20260902","catalog_created_at":"2026-09-02T15:14:16+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"google/gemini-3.8-flash:batch","input_price":null,"max_output_tokens":65536,"name":"Google: Gemini 3.8 Flash (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemini-3.8-flash-20260902","version_policy":"independent_snapshot","evaluation_references":[{"benchmark":"Vals Index","benchmark_version":"2.2","checked_at":"2026-10-26T00:20:45.595508+00:00","date":"2026-10-24","effort":"high","identity_evidence":{"checked_at":"2026-10-26T00:20:45.595508+00:00","effort":"high","release_date":"2026-09-02","sha256":"cb6d1f00c9e19f594bee14e69b6a72a9834895eb39d2d620567a5bfeda1fb2e5","source_name":"Gemini 3.8 Flash","url":"https://www.vals.ai/models/google_gemini-3.8-flash"},"kind":"independent","model_url":"https://www.vals.ai/models/google_gemini-3.8-flash","model_version":"google/gemini-3.8-flash-20260902","note":"按经济领域权重汇总的任务成绩；原榜含部分其他模型使用备用模型的结果。推理档位为来源模型页默认值，单项测试可能不同，完整条件见原榜。","rank":15,"scope":"该版本综合榜全部已评测模型","score":55.95,"source_id":"vals","source_model_id":"Gemini 3.8 Flash","source_url":"https://www.vals.ai/benchmarks/vals_index","total":27,"unit":"%","update_mode":"automatic","stale":false},{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":161.08,"ci_low":155.07,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-09-02"],"catalog_release_date":"2026-09-02","source_model_date":"2026-09-02","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"google/gemini-3.8-flash-20260902","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":20,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":157.44,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-09-02","source_model_id":"Gemini 3.8 Flash","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":54,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"google/gemma-4-26b-a4b-it-20260403","catalog_created_at":"2026-04-03T14:53:09+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"google/gemma-4-26b-a4b-it:free","input_price":0.0,"max_output_tokens":32768,"name":"Google: Gemma 4 26B A4B  (free)","new_model":false,"open_weights":true,"output_price":0.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2026-04-02","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemma-4-26b-a4b-it-20260403","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":207,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"google/gemma-4-31b-it-20260402","catalog_created_at":"2026-04-02T16:48:06+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"google/gemma-4-31b-it:free","input_price":0.0,"max_output_tokens":32768,"name":"Google: Gemma 4 31B (free)","new_model":false,"open_weights":true,"output_price":0.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"google","provider_label":"google","release_date":"2026-04-02","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"google/gemma-4-31b-it-20260402","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":207,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"liquid/lfm-2.5-2.6b-20260811","catalog_created_at":"2026-08-11T17:48:39+00:00","catalog_status":"active","context_length":65536,"eligible":false,"extra_pricing":{},"id":"liquid/lfm-2.5-2.6b:free","input_price":0.0,"max_output_tokens":8192,"name":"LiquidAI: LFM2.5-2.6B (free)","new_model":false,"open_weights":true,"output_price":0.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"liquid","provider_label":"liquid","release_date":"2026-08-11","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"liquid/lfm-2.5-2.6b-20260811","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":76,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"mistralai/codestral-2508","catalog_created_at":"2025-08-01T20:20:30+00:00","catalog_status":"active","context_length":256000,"eligible":false,"extra_pricing":{},"id":"mistralai/codestral-2508:batch","input_price":null,"max_output_tokens":204800,"name":"Mistral: Codestral 2508 (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"mistralai","provider_label":"mistralai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"mistralai/codestral-2508","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":451,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"mistralai/ministral-8b-2512","catalog_created_at":"2025-12-02T13:20:54+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"mistralai/ministral-8b-2512:batch","input_price":null,"max_output_tokens":209715,"name":"Mistral: Ministral 3 8B 2512 (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"mistralai","provider_label":"mistralai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"mistralai/ministral-8b-2512","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":328,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"mistralai/mistral-large-2512","catalog_created_at":"2025-12-01T21:27:52+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"mistralai/mistral-large-2512:batch","input_price":null,"max_output_tokens":209715,"name":"Mistral: Mistral Large 3 2512 (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"mistralai","provider_label":"mistralai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"mistralai/mistral-large-2512","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":329,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"mistralai/mistral-medium-3.1","catalog_created_at":"2025-08-13T14:33:59+00:00","catalog_status":"active","context_length":131072,"eligible":false,"extra_pricing":{},"id":"mistralai/mistral-medium-3.1:batch","input_price":null,"max_output_tokens":104857,"name":"Mistral: Mistral Medium 3.1 (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"mistralai","provider_label":"mistralai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"mistralai/mistral-medium-3.1","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":439,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"mistralai/mistral-medium-3.5-20260430","catalog_created_at":"2026-04-30T17:33:59+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"mistralai/mistral-medium-3-5:batch","input_price":null,"max_output_tokens":209715,"name":"Mistral: Mistral Medium 3.5 (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"mistralai","provider_label":"mistralai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"mistralai/mistral-medium-3.5-20260430","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":179,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"mistralai/mistral-small-2603","catalog_created_at":"2026-03-16T21:14:45+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"mistralai/mistral-small-2603:batch","input_price":null,"max_output_tokens":209715,"name":"Mistral: Mistral Small 4 (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"mistralai","provider_label":"mistralai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"mistralai/mistral-small-2603","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":224,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"moonshotai/kimi-k3-20260715","catalog_created_at":"2026-07-16T15:30:58+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"moonshotai/kimi-k3:batch","input_price":null,"max_output_tokens":16384,"name":"MoonshotAI: Kimi K3 (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"moonshotai","provider_label":"moonshotai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"moonshotai/kimi-k3-20260715","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":102,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning-20260428","catalog_created_at":"2026-04-28T16:18:15+00:00","catalog_status":"active","context_length":256000,"eligible":false,"extra_pricing":{},"id":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free","input_price":0.0,"max_output_tokens":65536,"name":"NVIDIA: Nemotron 3 Nano Omni (free)","new_model":false,"open_weights":true,"output_price":0.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"nvidia","provider_label":"nvidia","release_date":"2026-04-28","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"nvidia/nemotron-3-nano-omni-30b-a3b-reasoning-20260428","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":181,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"nvidia/nemotron-3-super-120b-a12b-20230311","catalog_created_at":"2026-03-11T16:07:19+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"nvidia/nemotron-3-super-120b-a12b:free","input_price":0.0,"max_output_tokens":235929,"name":"NVIDIA: Nemotron 3 Super (free)","new_model":false,"open_weights":true,"output_price":0.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"nvidia","provider_label":"nvidia","release_date":"2026-03-11","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"nvidia/nemotron-3-super-120b-a12b-20230311","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":229,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"nvidia/nemotron-3-ultra-550b-a55b-20260604","catalog_created_at":"2026-06-04T05:33:28+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"nvidia/nemotron-3-ultra-550b-a55b:free","input_price":0.0,"max_output_tokens":65536,"name":"NVIDIA: Nemotron 3 Ultra (free)","new_model":false,"open_weights":true,"output_price":0.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"nvidia","provider_label":"nvidia","release_date":"2026-06-04","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"nvidia/nemotron-3-ultra-550b-a55b-20260604","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":148.22,"ci_low":143.68,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-06-04"],"catalog_release_date":"2026-06-04","source_model_date":"2026-06-04","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"nvidia/nemotron-3-ultra-550b-a55b-20260604","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":84,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":145.97,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-06-04","source_model_id":"Nemotron 3 Ultra","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":144,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"nvidia/nemotron-3.5-content-safety-20260604","catalog_created_at":"2026-06-04T14:04:24+00:00","catalog_status":"active","context_length":128000,"eligible":false,"extra_pricing":{},"id":"nvidia/nemotron-3.5-content-safety:free","input_price":0.0,"max_output_tokens":8192,"name":"NVIDIA: Nemotron 3.5 Content Safety (free)","new_model":false,"open_weights":true,"output_price":0.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"nvidia","provider_label":"nvidia","release_date":"2026-06-02","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"nvidia/nemotron-3.5-content-safety-20260604","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":146,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"nvidia/nemotron-3.5-lightning-20260807","catalog_created_at":"2026-08-11T12:52:31+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"nvidia/nemotron-3.5-lightning:free","input_price":0.0,"max_output_tokens":65536,"name":"NVIDIA: Nemotron 3.5 Lightning (free)","new_model":false,"open_weights":true,"output_price":0.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"nvidia","provider_label":"nvidia","release_date":"2026-08-11","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"nvidia/nemotron-3.5-lightning-20260807","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":76,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-4-turbo","catalog_created_at":"2024-04-09T00:00:00+00:00","catalog_status":"active","context_length":128000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-4-turbo:batch","input_price":null,"max_output_tokens":4096,"name":"OpenAI: GPT-4 Turbo (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-4-turbo","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":930,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-4.1-2025-04-14","catalog_created_at":"2025-04-14T17:23:05+00:00","catalog_status":"active","context_length":1047576,"eligible":false,"extra_pricing":{},"id":"openai/gpt-4.1:batch","input_price":null,"max_output_tokens":32768,"name":"OpenAI: GPT-4.1 (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-4.1-2025-04-14","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":138.34,"ci_low":133.57,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-04-14"],"catalog_release_date":"2025-04-14","source_model_date":"2025-04-14","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-4.1-2025-04-14","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":137,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":136.65,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-04-14","source_model_id":"GPT-4.1","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":560,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-4.1-mini-2025-04-14","catalog_created_at":"2025-04-14T17:23:01+00:00","catalog_status":"active","context_length":1047576,"eligible":false,"extra_pricing":{},"id":"openai/gpt-4.1-mini:batch","input_price":null,"max_output_tokens":32768,"name":"OpenAI: GPT-4.1 Mini (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-4.1-mini-2025-04-14","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":136.49,"ci_low":130.97,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-04-14"],"catalog_release_date":"2025-04-14","source_model_date":"2025-04-14","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-4.1-mini-2025-04-14","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":145,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":134.89,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-04-14","source_model_id":"GPT-4.1 mini","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":560,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-4.1-nano-2025-04-14","catalog_created_at":"2025-04-14T17:22:49+00:00","catalog_status":"active","context_length":1047576,"eligible":false,"extra_pricing":{},"id":"openai/gpt-4.1-nano:batch","input_price":null,"max_output_tokens":32768,"name":"OpenAI: GPT-4.1 Nano (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-4.1-nano-2025-04-14","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":131.82,"ci_low":123.33,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-04-14"],"catalog_release_date":"2025-04-14","source_model_date":"2025-04-14","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-4.1-nano-2025-04-14","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":162,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":129.54,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-04-14","source_model_id":"GPT-4.1 nano","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":560,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-4o","catalog_created_at":"2024-05-13T00:00:00+00:00","catalog_status":"active","context_length":128000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-4o:batch","input_price":null,"max_output_tokens":16384,"name":"OpenAI: GPT-4o (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-4o","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":896,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-4o-mini","catalog_created_at":"2024-07-18T00:00:00+00:00","catalog_status":"active","context_length":128000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-4o-mini:batch","input_price":null,"max_output_tokens":16384,"name":"OpenAI: GPT-4o-mini (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-4o-mini","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":128.49,"ci_low":120.11,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":[],"catalog_release_date":"2024-07-18","source_model_date":"2024-07-18","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-4o-mini","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":177,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":126.46,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2024-07-18","source_model_id":"GPT-4o mini","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":830,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-5-2025-08-07","catalog_created_at":"2025-08-07T17:23:33+00:00","catalog_status":"active","context_length":400000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5:batch","input_price":null,"max_output_tokens":128000,"name":"OpenAI: GPT-5 (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5-2025-08-07","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":null,"ci_low":null,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-08-07"],"catalog_release_date":"2025-08-07","source_model_date":"2025-08-07","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-5-2025-08-07","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":61,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":150.0,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-08-07","source_model_id":"GPT-5","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":445,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-5-mini-2025-08-07","catalog_created_at":"2025-08-07T17:23:27+00:00","catalog_status":"active","context_length":400000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5-mini:batch","input_price":null,"max_output_tokens":128000,"name":"OpenAI: GPT-5 Mini (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5-mini-2025-08-07","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":147.11,"ci_low":143.33,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-08-07"],"catalog_release_date":"2025-08-07","source_model_date":"2025-08-07","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-5-mini-2025-08-07","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":89,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":145.34,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-08-07","source_model_id":"GPT-5 mini","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":445,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-5-nano-2025-08-07","catalog_created_at":"2025-08-07T17:23:22+00:00","catalog_status":"active","context_length":400000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5-nano:batch","input_price":null,"max_output_tokens":128000,"name":"OpenAI: GPT-5 Nano (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5-nano-2025-08-07","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":141.2,"ci_low":134.11,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-08-07"],"catalog_release_date":"2025-08-07","source_model_date":"2025-08-07","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-5-nano-2025-08-07","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":127,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":139.21,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-08-07","source_model_id":"GPT-5 nano","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":445,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-5-pro-2025-10-06","catalog_created_at":"2025-10-06T18:51:03+00:00","catalog_status":"active","context_length":400000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5-pro:batch","input_price":null,"max_output_tokens":128000,"name":"OpenAI: GPT-5 Pro (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5-pro-2025-10-06","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":385,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-5.1-20251113","catalog_created_at":"2025-11-13T18:58:25+00:00","catalog_status":"active","context_length":400000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5.1:batch","input_price":null,"max_output_tokens":128000,"name":"OpenAI: GPT-5.1 (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.1-20251113","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":151.74,"ci_low":148.36,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-11-13"],"catalog_release_date":"2025-11-13","source_model_date":"2025-11-13","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-5.1-20251113","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":63,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":149.76,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-11-13","source_model_id":"GPT-5.1","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":347,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-5.2-20251211","catalog_created_at":"2025-12-10T18:02:55+00:00","catalog_status":"active","context_length":400000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5.2:batch","input_price":null,"max_output_tokens":128000,"name":"OpenAI: GPT-5.2 (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.2-20251211","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":155.42,"ci_low":152.05,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-12-11"],"catalog_release_date":"2025-12-11","source_model_date":"2025-12-11","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-5.2-20251211","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":46,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":153.66,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-12-11","source_model_id":"GPT-5.2","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":320,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-5.2-pro-20251211","catalog_created_at":"2025-12-10T18:03:00+00:00","catalog_status":"active","context_length":400000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5.2-pro:batch","input_price":null,"max_output_tokens":128000,"name":"OpenAI: GPT-5.2 Pro (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.2-pro-20251211","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":159.98,"ci_low":153.89,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-12-11"],"catalog_release_date":"2025-12-11","source_model_date":"2025-12-11","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-5.2-pro-20251211","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":30,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":156.33,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-12-11","source_model_id":"GPT-5.2 Pro","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":320,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-5.4-20260305","catalog_created_at":"2026-03-05T18:12:32+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5.4:batch","input_price":null,"max_output_tokens":128000,"name":"OpenAI: GPT-5.4 (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.4-20260305","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":159.8,"ci_low":155.38,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-03-05"],"catalog_release_date":"2026-03-05","source_model_date":"2026-03-05","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-5.4-20260305","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":22,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":157.31,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-03-05","source_model_id":"GPT-5.4","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":235,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-5.4-mini-20260317","catalog_created_at":"2026-03-17T11:49:38+00:00","catalog_status":"active","context_length":400000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5.4-mini:batch","input_price":null,"max_output_tokens":128000,"name":"OpenAI: GPT-5.4 Mini (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.4-mini-20260317","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":150.47,"ci_low":146.87,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-03-17"],"catalog_release_date":"2026-03-17","source_model_date":"2026-03-17","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-5.4-mini-20260317","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":67,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":148.7,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-03-17","source_model_id":"GPT-5.4 Mini","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":223,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-5.4-nano-20260317","catalog_created_at":"2026-03-17T11:49:47+00:00","catalog_status":"active","context_length":400000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5.4-nano:batch","input_price":null,"max_output_tokens":128000,"name":"OpenAI: GPT-5.4 Nano (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.4-nano-20260317","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":148.15,"ci_low":143.63,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-03-17"],"catalog_release_date":"2026-03-17","source_model_date":"2026-03-17","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-5.4-nano-20260317","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":83,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":146.03,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-03-17","source_model_id":"GPT-5.4 Nano","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":223,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-5.4-pro-20260305","catalog_created_at":"2026-03-05T18:12:46+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5.4-pro:batch","input_price":null,"max_output_tokens":128000,"name":"OpenAI: GPT-5.4 Pro (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.4-pro-20260305","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":162.7,"ci_low":157.1,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-03-05"],"catalog_release_date":"2026-03-05","source_model_date":"2026-03-05","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-5.4-pro-20260305","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":15,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":159.57,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-03-05","source_model_id":"GPT-5.4 Pro","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":235,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-5.5-20260423","catalog_created_at":"2026-04-24T17:31:33+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5.5:batch","input_price":null,"max_output_tokens":128000,"name":"OpenAI: GPT-5.5 (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.5-20260423","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":163.67,"ci_low":157.33,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-04-23"],"catalog_release_date":"2026-04-23","source_model_date":"2026-04-23","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-5.5-20260423","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":14,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":159.68,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-04-23","source_model_id":"GPT-5.5","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":185,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-5.5-pro-20260423","catalog_created_at":"2026-04-24T17:31:36+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5.5-pro:batch","input_price":null,"max_output_tokens":128000,"name":"OpenAI: GPT-5.5 Pro (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.5-pro-20260423","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":166.6,"ci_low":160.04,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-04-23"],"catalog_release_date":"2026-04-23","source_model_date":"2026-04-23","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-5.5-pro-20260423","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":10,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":162.91,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-04-23","source_model_id":"GPT-5.5 Pro","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":185,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-5.6-luna-20260709","catalog_created_at":"2026-07-09T09:54:24+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5.6-luna:batch","input_price":null,"max_output_tokens":128000,"name":"OpenAI: GPT-5.6 Luna (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.6-luna-20260709","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":159.26,"ci_low":154.19,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-07-09"],"catalog_release_date":"2026-07-09","source_model_date":"2026-07-09","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-5.6-luna-20260709","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":29,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":156.5,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-07-09","source_model_id":"GPT-5.6 Luna","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":109,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-5.6-luna-pro-20260709","catalog_created_at":"2026-07-09T09:54:27+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5.6-luna-pro:batch","input_price":null,"max_output_tokens":128000,"name":"OpenAI: GPT-5.6 Luna Pro (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.6-luna-pro-20260709","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":109,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-5.6-sol-20260709","catalog_created_at":"2026-07-09T09:54:10+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5.6-sol:batch","input_price":null,"max_output_tokens":128000,"name":"OpenAI: GPT-5.6 Sol (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.6-sol-20260709","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":167.06,"ci_low":160.15,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-07-09"],"catalog_release_date":"2026-07-09","source_model_date":"2026-07-09","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-5.6-sol-20260709","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":11,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":162.79,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-07-09","source_model_id":"GPT-5.6 Sol","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":109,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-5.6-sol-pro-20260709","catalog_created_at":"2026-07-09T09:54:14+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5.6-sol-pro:batch","input_price":null,"max_output_tokens":128000,"name":"OpenAI: GPT-5.6 Sol Pro (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.6-sol-pro-20260709","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":109,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-5.6-terra-20260709","catalog_created_at":"2026-07-09T09:54:17+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5.6-terra:batch","input_price":null,"max_output_tokens":128000,"name":"OpenAI: GPT-5.6 Terra (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.6-terra-20260709","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":163.31,"ci_low":157.3,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-07-09"],"catalog_release_date":"2026-07-09","source_model_date":"2026-07-09","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-5.6-terra-20260709","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":13,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":160.09,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-07-09","source_model_id":"GPT-5.6 Terra","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":109,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-5.6-terra-pro-20260709","catalog_created_at":"2026-07-09T09:54:21+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-5.6-terra-pro:batch","input_price":null,"max_output_tokens":128000,"name":"OpenAI: GPT-5.6 Terra Pro (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-5.6-terra-pro-20260709","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":109,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-6-astra-20260903","catalog_created_at":"2026-09-04T20:13:58+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-6-astra:batch","input_price":null,"max_output_tokens":128000,"name":"OpenAI: GPT-6 Astra (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-6-astra-20260903","version_policy":"independent_snapshot","evaluation_references":[{"benchmark":"Vals Index","benchmark_version":"2.2","checked_at":"2026-10-26T00:20:45.595508+00:00","date":"2026-10-24","effort":"max","identity_evidence":{"checked_at":"2026-10-26T00:20:45.595508+00:00","effort":"max","release_date":"2026-09-03","sha256":"80f9fa8969e5eea06ab19566d34e685f47fdee7ba991c2436282601c462468e4","source_name":"GPT-6 Astra","url":"https://www.vals.ai/models/openai_gpt-6-astra"},"kind":"independent","model_url":"https://www.vals.ai/models/openai_gpt-6-astra","model_version":"openai/gpt-6-astra-20260903","note":"按经济领域权重汇总的任务成绩；原榜含部分其他模型使用备用模型的结果。推理档位为来源模型页默认值，单项测试可能不同，完整条件见原榜。","rank":5,"scope":"该版本综合榜全部已评测模型","score":65.02,"source_id":"vals","source_model_id":"GPT-6 Astra","source_url":"https://www.vals.ai/benchmarks/vals_index","total":27,"unit":"%","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":52,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-6-astra-pro-20260903","catalog_created_at":"2026-09-04T20:13:55+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-6-astra-pro:batch","input_price":null,"max_output_tokens":128000,"name":"OpenAI: GPT-6 Astra Pro (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-6-astra-pro-20260903","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":52,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-6-luna-20260922","catalog_created_at":"2026-09-22T18:13:06+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-6-luna:batch","input_price":null,"max_output_tokens":128000,"name":"OpenAI: GPT-6 Luna (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-6-luna-20260922","version_policy":"independent_snapshot","evaluation_references":[{"benchmark":"Vals Index","benchmark_version":"2.2","checked_at":"2026-10-26T00:20:45.595508+00:00","date":"2026-10-24","effort":"max","identity_evidence":{"checked_at":"2026-10-26T00:20:45.595508+00:00","effort":"max","release_date":"2026-09-22","sha256":"e483f5e76f48dd7922c8433af499f9d2b7fb71933f159096d90911fd7915c589","source_name":"GPT-6 Luna","url":"https://www.vals.ai/models/openai_gpt-6-luna"},"kind":"independent","model_url":"https://www.vals.ai/models/openai_gpt-6-luna","model_version":"openai/gpt-6-luna-20260922","note":"按经济领域权重汇总的任务成绩；原榜含部分其他模型使用备用模型的结果。推理档位为来源模型页默认值，单项测试可能不同，完整条件见原榜。","rank":21,"scope":"该版本综合榜全部已评测模型","score":52.32,"source_id":"vals","source_model_id":"GPT-6 Luna","source_url":"https://www.vals.ai/benchmarks/vals_index","total":27,"unit":"%","update_mode":"automatic","stale":false},{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":159.43,"ci_low":154.32,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-09-22"],"catalog_release_date":"2026-09-22","source_model_date":"2026-09-22","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-6-luna-20260922","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":27,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":156.75,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-09-22","source_model_id":"GPT-6 Luna","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":34,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-6-luna-pro-20260922","catalog_created_at":"2026-09-22T18:13:11+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-6-luna-pro:batch","input_price":null,"max_output_tokens":128000,"name":"OpenAI: GPT-6 Luna Pro (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-6-luna-pro-20260922","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":34,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-6-sol-20260922","catalog_created_at":"2026-09-22T18:12:55+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-6-sol:batch","input_price":null,"max_output_tokens":128000,"name":"OpenAI: GPT-6 Sol (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-6-sol-20260922","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":167.44,"ci_low":160.92,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-09-22"],"catalog_release_date":"2026-09-22","source_model_date":"2026-09-22","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-6-sol-20260922","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":8,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":163.73,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-09-22","source_model_id":"GPT-6 Sol","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":34,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-6-sol-pro-20260922","catalog_created_at":"2026-09-22T18:13:01+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"openai/gpt-6-sol-pro:batch","input_price":null,"max_output_tokens":128000,"name":"OpenAI: GPT-6 Sol Pro (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-6-sol-pro-20260922","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":34,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-oss-120b","catalog_created_at":"2025-08-05T17:17:11+00:00","catalog_status":"active","context_length":131072,"eligible":false,"extra_pricing":{},"id":"openai/gpt-oss-120b:batch","input_price":null,"max_output_tokens":117964,"name":"OpenAI: gpt-oss-120b (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-oss-120b","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":142.12,"ci_low":135.57,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":[],"catalog_release_date":"2025-08-05","source_model_date":"2025-08-05","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-oss-120b","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":123,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":139.79,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-08-05","source_model_id":"gpt-oss-120b","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":447,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/gpt-oss-20b","catalog_created_at":"2025-08-05T17:17:09+00:00","catalog_status":"active","context_length":131072,"eligible":false,"extra_pricing":{},"id":"openai/gpt-oss-20b:batch","input_price":null,"max_output_tokens":117964,"name":"OpenAI: gpt-oss-20b (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/gpt-oss-20b","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":139.58,"ci_low":132.2,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":[],"catalog_release_date":"2025-08-05","source_model_date":"2025-08-05","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/gpt-oss-20b","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":134,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":137.7,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-08-05","source_model_id":"gpt-oss-20b","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":447,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/o3-2025-04-16","catalog_created_at":"2025-04-16T17:10:57+00:00","catalog_status":"active","context_length":200000,"eligible":false,"extra_pricing":{},"id":"openai/o3:batch","input_price":null,"max_output_tokens":100000,"name":"OpenAI: o3 (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/o3-2025-04-16","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":148.52,"ci_low":144.83,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-04-16"],"catalog_release_date":"2025-04-16","source_model_date":"2025-04-16","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/o3-2025-04-16","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":76,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":146.73,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-04-16","source_model_id":"o3","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":558,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/o3-mini-2025-01-31","catalog_created_at":"2025-01-31T19:28:41+00:00","catalog_status":"active","context_length":200000,"eligible":false,"extra_pricing":{},"id":"openai/o3-mini:batch","input_price":null,"max_output_tokens":100000,"name":"OpenAI: o3 Mini (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/o3-mini-2025-01-31","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":633,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"openai/o4-mini-2025-04-16","catalog_created_at":"2025-04-16T16:29:02+00:00","catalog_status":"active","context_length":200000,"eligible":false,"extra_pricing":{},"id":"openai/o4-mini:batch","input_price":null,"max_output_tokens":100000,"name":"OpenAI: o4 Mini (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"openai","provider_label":"openai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"openai/o4-mini-2025-04-16","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":147.21,"ci_low":142.9,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2025-04-16"],"catalog_release_date":"2025-04-16","source_model_date":"2025-04-16","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"openai/o4-mini-2025-04-16","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":88,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":145.46,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2025-04-16","source_model_id":"o4-mini","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":558,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"poolside/laguna-s-2.1-20260720","catalog_created_at":"2026-07-21T16:51:23+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"poolside/laguna-s-2.1:free","input_price":0.0,"max_output_tokens":32768,"name":"Poolside: Laguna S 2.1 (free)","new_model":false,"open_weights":true,"output_price":0.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"poolside","provider_label":"poolside","release_date":"2026-07-21","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"poolside/laguna-s-2.1-20260720","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":97,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"poolside/laguna-xs-2.1-20260625","catalog_created_at":"2026-07-02T14:27:09+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"poolside/laguna-xs-2.1:free","input_price":0.0,"max_output_tokens":32768,"name":"Poolside: Laguna XS 2.1 (free)","new_model":false,"open_weights":true,"output_price":0.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"poolside","provider_label":"poolside","release_date":"2026-07-02","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"poolside/laguna-xs-2.1-20260625","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":116,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"qwen/qwen3.7-plus-20260602","catalog_created_at":"2026-06-03T13:03:03+00:00","catalog_status":"active","context_length":262144,"eligible":false,"extra_pricing":{},"id":"qwen/qwen3.7-plus:batch","input_price":null,"max_output_tokens":131072,"name":"Qwen: Qwen3.7 Plus (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"qwen","provider_label":"qwen","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"qwen/qwen3.7-plus-20260602","version_policy":"independent_snapshot","evaluation_references":[{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":148.82,"ci_low":145.25,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-06-02"],"catalog_release_date":"2026-06-02","source_model_date":"2026-06-02","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"qwen/qwen3.7-plus-20260602","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":72,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":147.0,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-06-02","source_model_id":"Qwen3.7-Plus","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":145,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"thinkingmachines/inkling-20260715","catalog_created_at":"2026-07-17T22:05:56+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"thinkingmachines/inkling:free","input_price":0.0,"max_output_tokens":262144,"name":"Thinking Machines: Inkling (free)","new_model":false,"open_weights":true,"output_price":0.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"thinkingmachines","provider_label":"thinkingmachines","release_date":"2026-07-15","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"thinkingmachines/inkling-20260715","version_policy":"independent_snapshot","evaluation_references":[{"benchmark":"Vals Index","benchmark_version":"2.2","checked_at":"2026-10-26T00:20:45.595508+00:00","date":"2026-10-24","effort":"0","identity_evidence":{"checked_at":"2026-10-26T00:20:45.595508+00:00","effort":"0","release_date":"2026-07-15","sha256":"e5869ecfaa80195bf0708f36d46938983202d53ba85d41ea648f3871369e80bb","source_name":"Inkling","url":"https://www.vals.ai/models/thinkingmachines_inkling"},"kind":"independent","model_url":"https://www.vals.ai/models/thinkingmachines_inkling","model_version":"thinkingmachines/inkling-20260715","note":"按经济领域权重汇总的任务成绩；原榜含部分其他模型使用备用模型的结果。推理档位为来源模型页默认值，单项测试可能不同，完整条件见原榜。","rank":26,"scope":"该版本综合榜全部已评测模型","score":28.96,"source_id":"vals","source_model_id":"Inkling","source_url":"https://www.vals.ai/benchmarks/vals_index","total":27,"unit":"%","update_mode":"automatic","stale":false},{"attribution":"Epoch AI, Epoch Capabilities Index（CC BY 4.0）；本网站按原始 ECI 值计算名次并匹配模型版本。","benchmark":"Epoch Capabilities Index","benchmark_version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","checked_at":"2026-10-26T00:20:48.326858+00:00","ci_high":150.21,"ci_low":145.96,"data_url":"https://epoch.ai/data/eci_scores.csv","date":"2026-10-26","date_kind":"observed_snapshot","effort":null,"identity_evidence":{"canonical_dates":["2026-07-15"],"catalog_release_date":"2026-07-15","source_model_date":"2026-07-15","source_version_ids":[]},"identity_match":"exact_name_provider_and_dates","kind":"independent_composite","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","model_version":"thinkingmachines/inkling-20260715","note":"ECI 汇总多个来源的基准结果，各基准的推理档位可能不同；未公布的统一档位不推断。置信区间按原源保留，ECI 分数不换算为 Arena Elo。","rank":68,"scope":"原始 CSV 全部已评分记录，含历史模型及源列出的配置；按 ECI 降序计算，并列同位，并非最新独立模型总榜。","score":148.41,"snapshot_date":"2026-10-26","source_id":"epoch","source_model_date":"2026-07-15","source_model_id":"Inkling","source_url":"https://epoch.ai/eci","total":270,"unit":"分","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":103,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"thinkingmachines/inkling-small-20260730","catalog_created_at":"2026-07-30T20:25:17+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"thinkingmachines/inkling-small:free","input_price":0.0,"max_output_tokens":262144,"name":"Thinking Machines: Inkling Small (free)","new_model":false,"open_weights":true,"output_price":0.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"thinkingmachines","provider_label":"thinkingmachines","release_date":"2026-07-30","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"thinkingmachines/inkling-small-20260730","version_policy":"independent_snapshot","evaluation_references":[{"benchmark":"Vals Index","benchmark_version":"2.2","checked_at":"2026-10-26T00:20:45.595508+00:00","date":"2026-10-24","effort":"0","identity_evidence":{"checked_at":"2026-10-26T00:20:45.595508+00:00","effort":"0","release_date":"2026-07-30","sha256":"a7081c15822570a1171b2f182a36141b9b5df44e864fb86990defbac93f73ce1","source_name":"Inkling Small","url":"https://www.vals.ai/models/thinkingmachines_inkling-small"},"kind":"independent","model_url":"https://www.vals.ai/models/thinkingmachines_inkling-small","model_version":"thinkingmachines/inkling-small-20260730","note":"按经济领域权重汇总的任务成绩；原榜含部分其他模型使用备用模型的结果。推理档位为来源模型页默认值，单项测试可能不同，完整条件见原榜。","rank":27,"scope":"该版本综合榜全部已评测模型","score":27.59,"source_id":"vals","source_model_id":"Inkling Small","source_url":"https://www.vals.ai/benchmarks/vals_index","total":27,"unit":"%","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":88,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"x-ai/grok-4.3-20260430","catalog_created_at":"2026-04-30T23:30:21+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"x-ai/grok-4.3:batch","input_price":null,"max_output_tokens":900000,"name":"SpaceXAI: Grok 4.3 (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"x-ai","provider_label":"x-ai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"x-ai/grok-4.3-20260430","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":179,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"z-ai/glm-5.3-20260816","catalog_created_at":"2026-08-18T20:57:35+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"z-ai/glm-5.3:batch","input_price":null,"max_output_tokens":131072,"name":"Z.ai: GLM 5.3 (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"z-ai","provider_label":"z-ai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"z-ai/glm-5.3-20260816","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":69,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":null,"canonical_slug":"z-ai/glm-5.3-flash-20260826","catalog_created_at":"2026-08-26T13:59:01+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"z-ai/glm-5.3-flash:batch","input_price":null,"max_output_tokens":131072,"name":"Z.ai: GLM 5.3 Flash (batch)","new_model":false,"open_weights":null,"output_price":null,"price_source":null,"price_updated_at":null,"pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"z-ai","provider_label":"z-ai","release_date":null,"score":null,"source_url":null,"unsupported_pricing_conditions":[],"price_stale":false,"model_version":"z-ai/glm-5.3-flash-20260826","version_policy":"independent_snapshot","evaluation_references":[{"benchmark":"Vals Index","benchmark_version":"2.2","checked_at":"2026-10-26T00:20:45.595508+00:00","date":"2026-10-24","effort":"max","identity_evidence":{"checked_at":"2026-10-26T00:20:45.595508+00:00","effort":"max","release_date":"2026-08-26","sha256":"fd790cd563635a77a0c10777b08f676910ce7e9b316b07e3f6ec975990661383","source_name":"GLM 5.3 Flash","url":"https://www.vals.ai/models/zai_glm-5.3-flash"},"kind":"independent","model_url":"https://www.vals.ai/models/zai_glm-5.3-flash","model_version":"z-ai/glm-5.3-flash-20260826","note":"按经济领域权重汇总的任务成绩；原榜含部分其他模型使用备用模型的结果。推理档位为来源模型页默认值，单项测试可能不同，完整条件见原榜。","rank":22,"scope":"该版本综合榜全部已评测模型","score":50.86,"source_id":"vals","source_model_id":"GLM 5.3 Flash","source_url":"https://www.vals.ai/benchmarks/vals_index","total":27,"unit":"%","update_mode":"automatic","stale":false}],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":61,"new_model_basis":"catalog_created_at","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.25,"canonical_slug":"~anthropic/claude-fable-latest","catalog_created_at":"2026-06-09T18:32:24+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"~anthropic/claude-fable-latest","input_price":10.0,"max_output_tokens":128000,"name":"Anthropic: Claude Fable Latest","new_model":false,"open_weights":false,"output_price":50.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"~anthropic","provider_label":"~anthropic","release_date":"2026-06-09","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"~anthropic/claude-fable-latest","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":139,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.01,"canonical_slug":"~anthropic/claude-haiku-latest","catalog_created_at":"2026-04-27T19:34:52+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"~anthropic/claude-haiku-latest","input_price":0.1,"max_output_tokens":128000,"name":"Anthropic: Claude Haiku Latest","new_model":false,"open_weights":false,"output_price":0.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"~anthropic","provider_label":"~anthropic","release_date":"2026-04-27","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"~anthropic/claude-haiku-latest","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":182,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.2,"canonical_slug":"~anthropic/claude-opus-latest","catalog_created_at":"2026-04-21T18:16:01+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"~anthropic/claude-opus-latest","input_price":4.0,"max_output_tokens":128000,"name":"Anthropic: Claude Opus Latest","new_model":false,"open_weights":false,"output_price":20.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"~anthropic","provider_label":"~anthropic","release_date":"2026-04-21","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"~anthropic/claude-opus-latest","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":188,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.1,"canonical_slug":"~anthropic/claude-sonnet-latest","catalog_created_at":"2026-04-27T19:32:48+00:00","catalog_status":"active","context_length":1000000,"eligible":false,"extra_pricing":{},"id":"~anthropic/claude-sonnet-latest","input_price":2.0,"max_output_tokens":128000,"name":"Anthropic: Claude Sonnet Latest","new_model":false,"open_weights":false,"output_price":10.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"~anthropic","provider_label":"~anthropic","release_date":"2026-04-27","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"~anthropic/claude-sonnet-latest","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":182,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.015,"canonical_slug":"~deepseek/deepseek-flash-latest","catalog_created_at":"2026-09-14T15:19:10+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"~deepseek/deepseek-flash-latest","input_price":0.02,"max_output_tokens":943718,"name":"DeepSeek: DeepSeek Flash Latest","new_model":false,"open_weights":false,"output_price":1.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"~deepseek","provider_label":"~deepseek","release_date":"2026-09-14","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"~deepseek/deepseek-flash-latest","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":42,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.12,"canonical_slug":"~deepseek/deepseek-pro-latest","catalog_created_at":"2026-09-14T15:19:34+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"~deepseek/deepseek-pro-latest","input_price":0.1188,"max_output_tokens":393216,"name":"DeepSeek: DeepSeek Pro Latest","new_model":false,"open_weights":false,"output_price":8.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"~deepseek","provider_label":"~deepseek","release_date":"2026-09-14","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"~deepseek/deepseek-pro-latest","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":42,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.027,"canonical_slug":"~deepseek/deepseek-v4-flash-latest","catalog_created_at":"2026-08-01T17:40:09+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"~deepseek/deepseek-v4-flash-latest","input_price":0.0422,"max_output_tokens":943718,"name":"DeepSeek: DeepSeek V4 Flash Latest","new_model":false,"open_weights":false,"output_price":0.1763,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"~deepseek","provider_label":"~deepseek","release_date":"2026-08-01","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"~deepseek/deepseek-v4-flash-latest","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":86,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.075,"canonical_slug":"~google/gemini-flash-latest","catalog_created_at":"2026-04-27T19:33:18+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"~google/gemini-flash-latest","input_price":0.75,"max_output_tokens":65536,"name":"Google: Gemini Flash Latest","new_model":false,"open_weights":false,"output_price":3.75,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"~google","provider_label":"~google","release_date":"2026-04-27","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"~google/gemini-flash-latest","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":182,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.1,"canonical_slug":"~google/gemini-pro-latest","catalog_created_at":"2026-04-27T19:34:11+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"~google/gemini-pro-latest","input_price":2.0,"max_output_tokens":943718,"name":"Google: Gemini Pro Latest","new_model":false,"open_weights":false,"output_price":10.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"~google","provider_label":"~google","release_date":"2026-04-27","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"~google/gemini-pro-latest","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":182,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.181,"canonical_slug":"~moonshotai/kimi-latest","catalog_created_at":"2026-04-27T19:33:48+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"~moonshotai/kimi-latest","input_price":0.191,"max_output_tokens":943718,"name":"MoonshotAI: Kimi Latest","new_model":false,"open_weights":false,"output_price":16.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"~moonshotai","provider_label":"~moonshotai","release_date":"2026-04-27","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"~moonshotai/kimi-latest","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":182,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":1.0,"canonical_slug":"~openai/gpt-astra-latest","catalog_created_at":"2026-09-11T12:48:52+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"~openai/gpt-astra-latest","input_price":10.0,"max_output_tokens":128000,"name":"OpenAI: GPT Astra Latest","new_model":false,"open_weights":false,"output_price":50.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"~openai","provider_label":"~openai","release_date":"2026-09-11","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"~openai/gpt-astra-latest","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":45,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.01,"canonical_slug":"~openai/gpt-luna-latest","catalog_created_at":"2026-09-11T12:48:42+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"~openai/gpt-luna-latest","input_price":0.1,"max_output_tokens":128000,"name":"OpenAI: GPT Luna Latest","new_model":false,"open_weights":false,"output_price":0.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"~openai","provider_label":"~openai","release_date":"2026-09-11","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"~openai/gpt-luna-latest","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":45,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.075,"canonical_slug":"~openai/gpt-mini-latest","catalog_created_at":"2026-04-27T19:34:31+00:00","catalog_status":"active","context_length":400000,"eligible":false,"extra_pricing":{},"id":"~openai/gpt-mini-latest","input_price":0.75,"max_output_tokens":128000,"name":"OpenAI: GPT Mini Latest","new_model":false,"open_weights":false,"output_price":4.5,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"~openai","provider_label":"~openai","release_date":"2026-04-27","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"~openai/gpt-mini-latest","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":182,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.1,"canonical_slug":"~openai/gpt-sol-latest","catalog_created_at":"2026-09-11T12:48:48+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"~openai/gpt-sol-latest","input_price":2.0,"max_output_tokens":128000,"name":"OpenAI: GPT Sol Latest","new_model":false,"open_weights":false,"output_price":10.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"~openai","provider_label":"~openai","release_date":"2026-09-11","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"~openai/gpt-sol-latest","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":45,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.2,"canonical_slug":"~openai/gpt-terra-latest","catalog_created_at":"2026-09-11T12:48:45+00:00","catalog_status":"active","context_length":1050000,"eligible":false,"extra_pricing":{},"id":"~openai/gpt-terra-latest","input_price":2.0,"max_output_tokens":128000,"name":"OpenAI: GPT Terra Latest","new_model":false,"open_weights":false,"output_price":12.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"~openai","provider_label":"~openai","release_date":"2026-09-11","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"~openai/gpt-terra-latest","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":45,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.5,"canonical_slug":"~x-ai/grok-latest","catalog_created_at":"2026-07-08T14:02:40+00:00","catalog_status":"active","context_length":500000,"eligible":false,"extra_pricing":{},"id":"~x-ai/grok-latest","input_price":2.0,"max_output_tokens":450000,"name":"xAI: Grok Latest","new_model":false,"open_weights":false,"output_price":6.0,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"~x-ai","provider_label":"~x-ai","release_date":"2026-07-08","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"~x-ai/grok-latest","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":110,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.03,"canonical_slug":"~z-ai/glm-flash-latest","catalog_created_at":"2026-08-27T08:00:33+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"~z-ai/glm-flash-latest","input_price":0.0513,"max_output_tokens":131072,"name":"Z.ai: GLM Flash Latest","new_model":false,"open_weights":false,"output_price":0.0701,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"~z-ai","provider_label":"~z-ai","release_date":"2026-08-27","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"~z-ai/glm-flash-latest","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":60,"new_model_basis":"release_date","variant":true},{"benchmark":null,"benchmark_date":null,"benchmark_source":null,"benchmark_status":"unscored","benchmark_version":null,"cache_read_price":0.0703,"canonical_slug":"~z-ai/glm-latest","catalog_created_at":"2026-08-19T14:50:53+00:00","catalog_status":"active","context_length":1048576,"eligible":false,"extra_pricing":{},"id":"~z-ai/glm-latest","input_price":0.0703,"max_output_tokens":131072,"name":"Z.ai: GLM Latest","new_model":false,"open_weights":false,"output_price":4.3843,"price_source":"models.dev","price_updated_at":"2026-10-26T00:20:44.736770+00:00","pricing_note":"models.dev 路由基础标价；未计算未收录的峰谷价、节假日价或供应商额外费用","pricing_tiers":[],"provider":"~z-ai","provider_label":"~z-ai","release_date":"2026-08-19","score":null,"source_url":"https://models.dev/catalog.json","unsupported_pricing_conditions":[],"price_stale":false,"model_version":"~z-ai/glm-latest","version_policy":"independent_snapshot","evaluation_references":[],"evaluation_history":[],"scored_scenarios":[],"effort":null,"mapping_type":null,"listed_days_ago":68,"new_model_basis":"release_date","variant":true}],"scenario":"overall","meta":{"updated_at":"2026-10-26T00:20:44.736770+00:00","price_updated_at":"2026-10-26T00:20:44.736770+00:00","benchmark_updated_at":"2026-10-25T23:26:05.356907+00:00","benchmark_publish_date":"2026-10-22","sources":[{"id":"openrouter","url":"https://openrouter.ai/api/v1/models","status":"ok","updated_at":"2026-10-26T00:20:44.000351+00:00","error":null,"count":467},{"id":"models.dev","url":"https://models.dev/catalog.json","status":"ok","updated_at":"2026-10-26T00:20:44.736770+00:00","error":null,"count":390},{"id":"arena","url":"https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset","license":"CC-BY-4.0","attribution":"Arena leaderboard data (lmarena-ai/leaderboard-dataset), CC BY 4.0. Model names mapped to OpenRouter routes by 斩杀线; mapping notes are published per model.","status":"ok","check_status":"ok","updated_at":"2026-10-25T23:26:05.356907+00:00","last_checked_at":"2026-10-26T00:20:50.587583+00:00","error":null,"count":1343,"revision":"9761a3ce414bac9ebac5a7742f67f230be809df4","publish_date":"2026-10-22","mapped_models":140,"source_models":418,"mapping_types":{"canonical":10,"date":10,"effort":35,"exact":85},"webdev_mapped":89,"last_attempt_at":null,"last_modified":"2026-10-24T03:02:41.000Z"},{"id":"deepseek_official","url":"https://api-docs.deepseek.com/quick_start/pricing/","status":"ok","updated_at":"2026-10-26T00:20:45.007063+00:00","error":null,"current_ids":["deepseek/deepseek-v4.1-flash-1017","deepseek/deepseek-v4-pro-0813"],"count":2},{"id":"vals","url":"https://www.vals.ai/benchmarks/vals_index","status":"ok","mapped_models":17,"unmatched_models":[{"candidates":["anthropic/claude-opus-5.5-20260921"],"model_url":"https://www.vals.ai/models/anthropic_claude-opus-5-5","reason":"release_date_unverified","source_model_id":"Claude Opus 5.5"},{"candidates":["google/gemini-4-argon-20261013"],"model_url":"https://www.vals.ai/models/google_gemini-4-argon","reason":"release_date_unverified","source_model_id":"Gemini 4 Argon"},{"candidates":[],"model_url":"https://www.vals.ai/models/meta_muse-wonder-full","reason":"no_catalog_identity","source_model_id":"Wonder Full"},{"candidates":[],"model_url":"https://www.vals.ai/models/meta_muse-wonder","reason":"no_catalog_identity","source_model_id":"Wonder"},{"candidates":["z-ai/glm-5.3-20260816"],"model_url":"https://www.vals.ai/models/zai_glm-5.3","reason":"release_date_unverified","source_model_id":"GLM 5.3"},{"candidates":["x-ai/grok-4.7-20260916"],"model_url":"https://www.vals.ai/models/grok_grok-4.7","reason":"release_date_unverified","source_model_id":"Grok 4.7"},{"candidates":["minimax/minimax-m3.1-flash-20261019"],"model_url":"https://www.vals.ai/models/minimax_MiniMax-M3.1-flash","reason":"release_date_unverified","source_model_id":"MiniMax-M3.1-Flash"},{"candidates":["tencent/hy4-preview-20260827"],"model_url":"https://www.vals.ai/models/tencent_hy4-preview","reason":"release_date_unverified","source_model_id":"Hy4 Preview"},{"candidates":["moonshotai/kimi-k3-20260715"],"model_url":"https://www.vals.ai/models/kimi_kimi-k3","reason":"release_date_unverified","source_model_id":"Kimi K3"},{"candidates":[],"model_url":"https://www.vals.ai/models/reflection_Beam-1.0","reason":"no_catalog_identity","source_model_id":"Beam 1.0"}],"updated_at":"2026-10-26T00:20:45.595508+00:00","error":null,"version":"2.2","publish_date":"2026-10-24","source_models":27,"cache_age_seconds":101386,"http_etag":"W/\"79de01778e0dec9bc3f832b2338654b8\"","http_last_modified":"Sat, 24 Oct 2026 20:10:58 GMT","evidence_sha256":"467b2e4868a44d9224bd1ad83c81aabed1e25a7a36eeeada681470e809198eb0","identity_errors":{},"last_attempt_at":null},{"id":"epoch","status":"ok","url":"https://epoch.ai/data/eci_scores.csv","license":"CC-BY-4.0","attribution":"Epoch AI, Epoch Capabilities Index, CC BY 4.0. Scores and ranks are independent of Arena.","mapped_models":73,"updated_at":"2026-10-26T00:20:48.326858+00:00","error":null,"version":"CSV SHA-256:540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","revision":"6eb82acf428d625c411b09dde93209b50731a4e3e3063416d9d3999c91fadf34","source_models":270,"evidence_sha256":"540796e67e6158f290edfb97a2633ca871edf9f1a89c78fc9d06cd20ca33cf64","unmatched_models":[{"candidate_versions":["anthropic/claude-opus-5.5-20260921"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2026-09-21"],"catalog_release_date":"2026-09-22","model_version":"anthropic/claude-opus-5.5-20260921","reason":"version_date_conflict","source_model_date":"2026-09-22"}],"source_model_date":"2026-09-22","source_model_id":"Claude Opus 5.5"},{"candidate_versions":["openai/gpt-6-astra-20260903"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2026-09-03"],"catalog_release_date":"2026-09-04","model_version":"openai/gpt-6-astra-20260903","reason":"version_date_conflict","source_model_date":"2026-09-03"}],"source_model_date":"2026-09-03","source_model_id":"GPT-6 Astra"},{"candidate_versions":["anthropic/claude-fable-5.1-20260831"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2026-08-31"],"catalog_release_date":"2026-09-01","model_version":"anthropic/claude-fable-5.1-20260831","reason":"version_date_conflict","source_model_date":"2026-09-01"}],"source_model_date":"2026-09-01","source_model_id":"Claude Fable 5.1"},{"candidate_versions":["google/gemini-4-argon-20261013"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2026-10-13"],"catalog_release_date":"2026-10-13","model_version":"google/gemini-4-argon-20261013","reason":"version_date_conflict","source_model_date":"2026-09-30"}],"source_model_date":"2026-09-30","source_model_id":"Gemini 4 Argon"},{"candidate_versions":["anthropic/claude-opus-5-20260723"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2026-07-23"],"catalog_release_date":"2026-07-24","model_version":"anthropic/claude-opus-5-20260723","reason":"version_date_conflict","source_model_date":"2026-07-24"}],"source_model_date":"2026-07-24","source_model_id":"Claude Opus 5"},{"candidate_versions":["moonshotai/kimi-k3-20260715"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2026-07-15"],"catalog_release_date":"2026-07-16","model_version":"moonshotai/kimi-k3-20260715","reason":"version_date_conflict","source_model_date":"2026-07-16"}],"source_model_date":"2026-07-16","source_model_id":"Kimi K3"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2026-08-02","source_model_id":"Qwen 3.8 Max"},{"candidate_versions":["openai/gpt-5.3-codex-20260224"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2026-02-24"],"catalog_release_date":"2026-02-05","model_version":"openai/gpt-5.3-codex-20260224","reason":"version_date_conflict","source_model_date":"2026-02-05"}],"source_model_date":"2026-02-05","source_model_id":"GPT-5.3 Codex"},{"candidate_versions":["x-ai/grok-4.6-20260810"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2026-08-10"],"catalog_release_date":"2026-08-12","model_version":"x-ai/grok-4.6-20260810","reason":"version_date_conflict","source_model_date":"2026-08-12"}],"source_model_date":"2026-08-12","source_model_id":"Grok 4.6"},{"candidate_versions":["z-ai/glm-5.3-20260816"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2026-08-16"],"catalog_release_date":"2026-08-14","model_version":"z-ai/glm-5.3-20260816","reason":"version_date_conflict","source_model_date":"2026-08-14"}],"source_model_date":"2026-08-14","source_model_id":"GLM-5.3"},{"candidate_versions":["deepseek/deepseek-v4.1-flash-20260910"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2026-09-10"],"catalog_release_date":"2026-09-10","model_version":"deepseek/deepseek-v4.1-flash-20260910","reason":"version_date_conflict","source_model_date":"2026-09-09"}],"source_model_date":"2026-09-09","source_model_id":"DeepSeek V4.1 Flash"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2026-02-19","source_model_id":"Gemini 3.1 Pro"},{"candidate_versions":["qwen/qwen3.8-max-20260902"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2026-09-02"],"catalog_release_date":"2026-09-02","model_version":"qwen/qwen3.8-max-20260902","reason":"version_date_conflict","source_model_date":"2026-09-01"}],"source_model_date":"2026-09-01","source_model_id":"Qwen3.8 Max (0902)"},{"candidate_versions":["x-ai/grok-4.7-20260916"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2026-09-16"],"catalog_release_date":"2026-09-21","model_version":"x-ai/grok-4.7-20260916","reason":"version_date_conflict","source_model_date":"2026-09-21"}],"source_model_date":"2026-09-21","source_model_id":"Grok 4.7"},{"candidate_versions":["meta/muse-spark-1.1-20260709"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2026-07-09"],"catalog_release_date":"2026-04-08","model_version":"meta/muse-spark-1.1-20260709","reason":"version_date_conflict","source_model_date":"2026-07-09"}],"source_model_date":"2026-07-09","source_model_id":"Muse Spark 1.1"},{"candidate_versions":["qwen/qwen3.7-max-20260520"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2026-05-20"],"catalog_release_date":"2026-05-21","model_version":"qwen/qwen3.7-max-20260520","reason":"version_date_conflict","source_model_date":"2026-05-19"}],"source_model_date":"2026-05-19","source_model_id":"Qwen3.7-Max"},{"candidate_versions":["z-ai/glm-5.3-flash-20260826"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2026-08-26"],"catalog_release_date":"2026-08-26","model_version":"z-ai/glm-5.3-flash-20260826","reason":"version_date_conflict","source_model_date":"2026-08-20"}],"source_model_date":"2026-08-20","source_model_id":"GLM-5.3-Flash"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-11-18","source_model_id":"Gemini 3 Pro"},{"candidate_versions":["x-ai/grok-4.20-20260309"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2026-03-09"],"catalog_release_date":"2026-03-31","model_version":"x-ai/grok-4.20-20260309","reason":"version_date_conflict","source_model_date":"2026-02-17"}],"source_model_date":"2026-02-17","source_model_id":"Grok 4.20"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2026-04-08","source_model_id":"Muse Spark"},{"candidate_versions":["z-ai/glm-5.2-20260616"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2026-06-16"],"catalog_release_date":"2026-06-13","model_version":"z-ai/glm-5.2-20260616","reason":"version_date_conflict","source_model_date":"2026-06-16"}],"source_model_date":"2026-06-16","source_model_id":"GLM-5.2"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-12-17","source_model_id":"Gemini 3 Flash"},{"candidate_versions":["openai/gpt-5-pro-2025-10-06"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2025-10-06"],"catalog_release_date":"2025-10-06","model_version":"openai/gpt-5-pro-2025-10-06","reason":"version_date_conflict","source_model_date":"2025-10-07"}],"source_model_date":"2025-10-07","source_model_id":"GPT-5 Pro"},{"candidate_versions":["thinkingmachines/inkling-small-20260730"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2026-07-30"],"catalog_release_date":"2026-07-30","model_version":"thinkingmachines/inkling-small-20260730","reason":"version_date_conflict","source_model_date":"2026-07-15"}],"source_model_date":"2026-07-15","source_model_id":"Inkling-Small"},{"candidate_versions":["z-ai/glm-5.1-20260406"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2026-04-06"],"catalog_release_date":"2026-04-07","model_version":"z-ai/glm-5.1-20260406","reason":"version_date_conflict","source_model_date":"2026-04-07"}],"source_model_date":"2026-04-07","source_model_id":"GLM-5.1"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2026-04-20","source_model_id":"Qwen 3.6 Max (Preview)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2026-04-24","source_model_id":"DeepSeek-V4-Pro"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2026-04-17","source_model_id":"Grok 4.3 Beta"},{"candidate_versions":["qwen/qwen3.6-plus-04-02"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":[],"catalog_release_date":"2026-04-02","model_version":"qwen/qwen3.6-plus-04-02","reason":"version_date_conflict","source_model_date":"2026-03-31"}],"source_model_date":"2026-03-31","source_model_id":"Qwen 3.6 Plus"},{"candidate_versions":["minimax/minimax-m3-20260531"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2026-05-31"],"catalog_release_date":"2026-06-01","model_version":"minimax/minimax-m3-20260531","reason":"version_date_conflict","source_model_date":"2026-06-01"}],"source_model_date":"2026-06-01","source_model_id":"MiniMax-M3"},{"candidate_versions":["qwen/qwen3.5-397b-a17b-20260216"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2026-02-16"],"catalog_release_date":"2026-02-15","model_version":"qwen/qwen3.5-397b-a17b-20260216","reason":"version_date_conflict","source_model_date":"2026-02-13"}],"source_model_date":"2026-02-13","source_model_id":"Qwen3.5 397B-A17B"},{"candidate_versions":["minimax/minimax-m2.5-20260211"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2026-02-11"],"catalog_release_date":"2026-02-12","model_version":"minimax/minimax-m2.5-20260211","reason":"version_date_conflict","source_model_date":"2026-02-12"}],"source_model_date":"2026-02-12","source_model_id":"MiniMax-M2.5"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-07-09","source_model_id":"Grok 4"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2026-02-16","source_model_id":"Qwen 3.5 Plus (hosted 397B-A17B)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2026-04-24","source_model_id":"DeepSeek-V4-Flash"},{"candidate_versions":["z-ai/glm-5-20260211"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2026-02-11"],"catalog_release_date":"2026-02-12","model_version":"z-ai/glm-5-20260211","reason":"version_date_conflict","source_model_date":"2026-02-11"}],"source_model_date":"2026-02-11","source_model_id":"GLM-5"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-06-05","source_model_id":"Gemini 2.5 Pro (Jun 2025)"},{"candidate_versions":["google/gemini-3.1-flash-lite-20260507"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2026-05-07"],"catalog_release_date":"2026-05-07","model_version":"google/gemini-3.1-flash-lite-20260507","reason":"version_date_conflict","source_model_date":"2026-03-03"}],"source_model_date":"2026-03-03","source_model_id":"Gemini 3.1 Flash-Lite"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-09-19","source_model_id":"Grok 4 Fast"},{"candidate_versions":["qwen/qwen3.7-flash-20260727"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2026-07-27"],"catalog_release_date":"2026-07-15","model_version":"qwen/qwen3.7-flash-20260727","reason":"version_date_conflict","source_model_date":"2026-07-27"}],"source_model_date":"2026-07-27","source_model_id":"Qwen3.7 Flash"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-03-31","source_model_id":"Gemini 2.5 Pro (Mar 2025)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-07-25","source_model_id":"Qwen3-235B-A22B-Thinking (Jul 2025)"},{"candidate_versions":["qwen/qwen3.6-35b-a3b-20260415"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2026-04-15"],"catalog_release_date":"2026-04-17","model_version":"qwen/qwen3.6-35b-a3b-20260415","reason":"version_date_conflict","source_model_date":"2026-04-14"}],"source_model_date":"2026-04-14","source_model_id":"Qwen 3.6 35B-A3B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2026-02-25","source_model_id":"Qwen 3.5 Flash (hosted 35B-A3B)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2026-04-02","source_model_id":"Gemma 4 31B IT"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-09-25","source_model_id":"Gemini 2.5 Flash (Sep 2025)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-05-22","source_model_id":"Claude Opus 4"},{"candidate_versions":["qwen/qwen3.5-35b-a3b-20260224"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2026-02-24"],"catalog_release_date":"2026-02-23","model_version":"qwen/qwen3.5-35b-a3b-20260224","reason":"version_date_conflict","source_model_date":"2026-02-24"}],"source_model_date":"2026-02-24","source_model_id":"Qwen3.5-35B-A3B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2026-05-05","source_model_id":"GPT-5.5 Instant"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-09-24","source_model_id":"Qwen3-Max"},{"candidate_versions":["anthropic/claude-4.5-haiku-20251001"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2025-10-01"],"catalog_release_date":"2025-10-15","model_version":"anthropic/claude-4.5-haiku-20251001","reason":"version_date_conflict","source_model_date":"2025-10-15"}],"source_model_date":"2025-10-15","source_model_id":"Claude Haiku 4.5"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-05-06","source_model_id":"Gemini 2.5 Pro (May 2025)"},{"candidate_versions":["google/gemma-4-26b-a4b-it-20260403"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2026-04-03"],"catalog_release_date":"2026-04-02","model_version":"google/gemma-4-26b-a4b-it-20260403","reason":"version_date_conflict","source_model_date":"2026-04-02"}],"source_model_date":"2026-04-02","source_model_id":"Gemma 4 26B A4B"},{"candidate_versions":["openai/o1-2024-12-17"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2024-12-17"],"catalog_release_date":"2024-12-05","model_version":"openai/o1-2024-12-17","reason":"version_date_conflict","source_model_date":"2024-12-17"}],"source_model_date":"2024-12-17","source_model_id":"o1"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-05-22","source_model_id":"Claude Sonnet 4"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-05-20","source_model_id":"Gemini 2.5 Flash (May 2025)"},{"candidate_versions":["mistralai/mistral-medium-3.5-20260430"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2026-04-30"],"catalog_release_date":"2026-04-30","model_version":"mistralai/mistral-medium-3.5-20260430","reason":"version_date_conflict","source_model_date":"2026-04-28"}],"source_model_date":"2026-04-28","source_model_id":"Mistral Medium 3.5"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-05-28","source_model_id":"DeepSeek-R1 (May 2025)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-02-24","source_model_id":"Claude 3.7 Sonnet"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-06-17","source_model_id":"Gemini 2.5 Flash (Jun 2025)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-06-24","source_model_id":"Grok-3 mini"},{"candidate_versions":["openai/o3-mini-2025-01-31"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2025-01-31"],"catalog_release_date":"2024-12-20","model_version":"openai/o3-mini-2025-01-31","reason":"version_date_conflict","source_model_date":"2025-01-31"}],"source_model_date":"2025-01-31","source_model_id":"o3-mini"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-07-12","source_model_id":"Kimi K2 (Jul 2025)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-04-17","source_model_id":"Gemini 2.5 Flash (Apr 2025)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-07-30","source_model_id":"Qwen3-30B-A3B-Thinking (Jul 2025)"},{"candidate_versions":["qwen/qwen3.5-9b-20260310"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":["2026-03-10"],"catalog_release_date":"2026-02-23","model_version":"qwen/qwen3.5-9b-20260310","reason":"version_date_conflict","source_model_date":"2026-02-24"}],"source_model_date":"2026-02-24","source_model_id":"Qwen3.5-9B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-04-28","source_model_id":"Qwen3-235B-A22B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-01-20","source_model_id":"DeepSeek-R1"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-07-25","source_model_id":"Qwen3-235B-A22B-Instruct (Jul 2025)"},{"candidate_versions":["qwen/qwen3-32b-04-28"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":[],"catalog_release_date":"2025-04-28","model_version":"qwen/qwen3-32b-04-28","reason":"version_date_conflict","source_model_date":"2025-04-29"}],"source_model_date":"2025-04-29","source_model_id":"Qwen3-32B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-04-09","source_model_id":"Grok 3"},{"candidate_versions":["qwen/qwen3-14b-04-28"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":[],"catalog_release_date":"2025-04-28","model_version":"qwen/qwen3-14b-04-28","reason":"version_date_conflict","source_model_date":"2025-04-29"}],"source_model_date":"2025-04-29","source_model_id":"Qwen3-14B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-03-05","source_model_id":"QwQ-32B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-07-29","source_model_id":"Qwen3-30B-A3B-Instruct (Jul 2025)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-02-27","source_model_id":"GPT-4.5"},{"candidate_versions":["qwen/qwen3-30b-a3b-04-28"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":[],"catalog_release_date":"2025-04-28","model_version":"qwen/qwen3-30b-a3b-04-28","reason":"version_date_conflict","source_model_date":"2025-04-29"}],"source_model_date":"2025-04-29","source_model_id":"Qwen3-30B-A3B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-04-28","source_model_id":"Qwen3-8B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-03-24","source_model_id":"DeepSeek-V3 (Mar 2025)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-09-12","source_model_id":"o1-mini"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-01-21","source_model_id":"Gemini 2.0 Flash Thinking (Jan 2025)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-02-05","source_model_id":"Gemini 2.0 Pro"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-02-05","source_model_id":"Gemini 2.0 Flash (Feb 2025)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-12-11","source_model_id":"Gemini 2.0 Flash (Dec 2024)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-09-12","source_model_id":"o1-preview"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-06-17","source_model_id":"Gemini 2.5 Flash-Lite (Jun 2025)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-10-22","source_model_id":"Claude 3.5 Sonnet (October 2024)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-06-10","source_model_id":"Magistral Small 1.0"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-01-25","source_model_id":"Qwen2.5-Max"},{"candidate_versions":["deepseek/deepseek-chat-v3"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":[],"catalog_release_date":"2025-12-01","model_version":"deepseek/deepseek-chat-v3","reason":"version_date_conflict","source_model_date":"2024-12-26"}],"source_model_date":"2024-12-26","source_model_id":"DeepSeek-V3"},{"candidate_versions":["meta-llama/llama-4-maverick-17b-128e-instruct"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":[],"catalog_release_date":"2025-04-05","model_version":"meta-llama/llama-4-maverick-17b-128e-instruct","reason":"version_date_conflict","source_model_date":"2025-04-06"}],"source_model_date":"2025-04-06","source_model_id":"Llama 4 Maverick"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-09-24","source_model_id":"Gemini 1.5 Pro (Sept 2024)"},{"candidate_versions":["microsoft/phi-4"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":[],"catalog_release_date":"2025-01-10","model_version":"microsoft/phi-4","reason":"version_date_conflict","source_model_date":"2024-12-12"}],"source_model_date":"2024-12-12","source_model_id":"Phi-4"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-12-12","source_model_id":"Grok-2 (Dec 2024)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-06-20","source_model_id":"Claude 3.5 Sonnet"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-09-24","source_model_id":"Gemini 1.5 Flash (Sep 2024)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-05-13","source_model_id":"GPT-4o (May 2024)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-09-19","source_model_id":"Qwen2.5-72B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-08-06","source_model_id":"GPT-4o (Aug 2024)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-07-23","source_model_id":"Llama 3.1-405B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-11-20","source_model_id":"GPT-4o (Nov 2024)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-11-18","source_model_id":"Mistral Large 2 (Nov 2024)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-07-24","source_model_id":"Mistral Large 2 (Jul 2024)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2025-03-17","source_model_id":"Mistral Small 3.1"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-12-06","source_model_id":"Llama 3.3 70B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-04-09","source_model_id":"GPT-4 Turbo (Apr 2024)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-10-22","source_model_id":"Claude 3.5 Haiku"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-05-14","source_model_id":"Gemini 1.5 Pro (May 2024)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-02-29","source_model_id":"Claude 3 Opus"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-01-25","source_model_id":"GPT-4 Turbo (Nov 2023)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-07-23","source_model_id":"Llama 3.1-70B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-03-14","source_model_id":"GPT-4 (Mar 2023)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-09-24","source_model_id":"Llama 3.2 90B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-06-07","source_model_id":"Qwen2-72B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-05-07","source_model_id":"DeepSeek-V2 (MoE-236B, May 2024)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-06-13","source_model_id":"GPT-4 (Jun 2023)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-04-18","source_model_id":"Llama 3-70B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-05-23","source_model_id":"Gemini 1.5 Flash (May 2024)"},{"candidate_versions":["google/gemma-2-27b-it"],"model_version":null,"reason":"version_date_conflict","rejected_candidates":[{"canonical_dates":[],"catalog_release_date":"2024-07-13","model_version":"google/gemma-2-27b-it","reason":"version_date_conflict","source_model_date":"2024-06-24"}],"source_model_date":"2024-06-24","source_model_id":"Gemma 2 27B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-04-17","source_model_id":"Mixtral 8x22B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-04-23","source_model_id":"phi-3-small 7.4B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-04-23","source_model_id":"phi-3-medium 14B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-02-29","source_model_id":"Claude 3 Sonnet"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-08-09","source_model_id":"Claude Instant"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-07-11","source_model_id":"Claude 2"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-06-24","source_model_id":"Gemma 2 9B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-09-18","source_model_id":"Qwen2.5-Coder-32B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-11-21","source_model_id":"Claude 2.1"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-08-30","source_model_id":"Command R+"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-03-07","source_model_id":"Claude 3 Haiku"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-09-19","source_model_id":"Qwen2.5-7B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-11-06","source_model_id":"GPT-3.5 Turbo (Nov 2023)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-12-11","source_model_id":"Mixtral 8x7B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-10-16","source_model_id":"Ministral 3B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-11-02","source_model_id":"Yi-34B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-04-23","source_model_id":"phi-3-mini 3.8B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-12-13","source_model_id":"Gemini 1.0 Pro"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-07-20","source_model_id":"Stable Beluga 2"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-07-23","source_model_id":"Llama 3.1-8B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-04-18","source_model_id":"Llama 3-8B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-09-18","source_model_id":"Qwen2.5-Coder-14B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-01-25","source_model_id":"GPT-3.5 Turbo (Jan 2024)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-05-17","source_model_id":"PaLM 2-L"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-07-18","source_model_id":"Llama 2-70B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-06-13","source_model_id":"GPT-3.5 Turbo (Jun 2023)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-09-18","source_model_id":"Qwen2.5-Coder (7B)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-09-24","source_model_id":"Qwen-14B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-09-27","source_model_id":"Mistral 7B v0.1"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-09-06","source_model_id":"Falcon-180B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-09-18","source_model_id":"internlm-20b"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-02-21","source_model_id":"Gemma 7B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-02-24","source_model_id":"LLaMA-65B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-05-09","source_model_id":"Falcon 2 11B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-05-27","source_model_id":"Mistral 7B v0.3"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-06-13","source_model_id":"DeepSeek-Coder-V2-Lite-Base"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-05-17","source_model_id":"PaLM 2-M"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-12-12","source_model_id":"Phi-2"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-09-18","source_model_id":"Qwen2.5-Coder-3B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-02-26","source_model_id":"Nemotron-4 15B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-03-01","source_model_id":"Yi-9B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-02-24","source_model_id":"LLaMA-33B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-09-28","source_model_id":"Qwen-7B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-05-17","source_model_id":"PaLM 2-S"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-07-18","source_model_id":"Llama 2-13B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-07-18","source_model_id":"Llama 2-34B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-02-20","source_model_id":"StarCoder 2 15B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-11-22","source_model_id":"Yi 6B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-05-25","source_model_id":"Falcon-40B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-09-06","source_model_id":"Baichuan2-13B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-09-18","source_model_id":"Qwen2.5-Coder (1.5B)"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-07-05","source_model_id":"internlm-7b"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-11-29","source_model_id":"INTELLECT-1"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-06-22","source_model_id":"MPT-30B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-02-24","source_model_id":"LLaMA-13B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-07-18","source_model_id":"Llama 2-7B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-06-24","source_model_id":"chatglm2-6b"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-02-24","source_model_id":"LLaMA-7B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-09-20","source_model_id":"Baichuan 2-7B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-11-02","source_model_id":"DeepSeek Coder 33B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-04-24","source_model_id":"Falcon-7B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-04-15","source_model_id":"CodeQwen1.5-7B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-04-12","source_model_id":"vicuna-13b-v1.1"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-05-05","source_model_id":"MPT-7B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-02-21","source_model_id":"Gemma 2B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-02-20","source_model_id":"StarCoder 2 7B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-06-27","source_model_id":"XGen-7B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-11-30","source_model_id":"Qwen-1_8B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-06-07","source_model_id":"open_llama_7b"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-09-11","source_model_id":"Phi-1.5"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-06-01","source_model_id":"Baichuan1-7B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-05-04","source_model_id":"RedPajama-INCITE-7B-Base"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-04-11","source_model_id":"Dolly 2.0-12b"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-11-02","source_model_id":"DeepSeek Coder 6.7B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-02-22","source_model_id":"StarCoder 2 3B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2024-09-18","source_model_id":"Qwen2.5-Coder-0.5B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-03-20","source_model_id":"Cerebras-GPT-13B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-11-02","source_model_id":"DeepSeek Coder 1.3B"},{"candidate_versions":[],"model_version":null,"reason":"no_exact_catalog_identity","source_model_date":"2023-04-19","source_model_id":"stablelm-tuned-alpha-7b"}],"last_attempt_at":null}],"refresh_status":{"arena":{"revision":"9761a3ce414bac9ebac5a7742f67f230be809df4","status":"unchanged"},"arena_error":null,"catalog":{"added":0,"removed":0,"total":467},"errors":{},"events":0,"finished_at":"2026-10-26T00:20:50.701070+00:00","status":"ok"},"scenarios":[{"id":"overall","label":"综合","hint":"日常问答、总结、综合任务","source_label":"Arena 文本榜 · 综合","benchmark":"arena/text_style_control/overall","scored":140},{"id":"coding","label":"写代码","hint":"写代码、改代码、解释代码","source_label":"Arena 文本榜 · 编程","benchmark":"arena/text_style_control/coding","scored":140},{"id":"webdev","label":"做网页","hint":"生成可运行的网页和前端界面","source_label":"Arena 网页开发榜","benchmark":"arena/webdev/overall","scored":89},{"id":"chinese","label":"中文","hint":"中文提问与中文写作","source_label":"Arena 文本榜 · 中文","benchmark":"arena/text_style_control/chinese","scored":137},{"id":"creative_writing","label":"写作","hint":"文案、故事、创意写作","source_label":"Arena 文本榜 · 创意写作","benchmark":"arena/text_style_control/creative_writing","scored":140},{"id":"math","label":"数学","hint":"数学与推理题","source_label":"Arena 文本榜 · 数学","benchmark":"arena/text_style_control/math","scored":137},{"id":"hard_prompts","label":"复杂难题","hint":"多步骤、要求高的复杂提问","source_label":"Arena 文本榜 · 难题","benchmark":"arena/text_style_control/hard_prompts","scored":140},{"id":"instruction_following","label":"按格式办事","hint":"严格按要求输出、信息抽取","source_label":"Arena 文本榜 · 指令遵循","benchmark":"arena/text_style_control/instruction_following","scored":140},{"id":"longer_query","label":"长文本","hint":"输入很长的文档或资料","source_label":"Arena 文本榜 · 长输入","benchmark":"arena/text_style_control/longer_query","scored":140},{"id":"multi_turn","label":"多轮对话","hint":"来回多轮的对话","source_label":"Arena 文本榜 · 多轮","benchmark":"arena/text_style_control/multi_turn","scored":140}],"default_scenario":"overall","counts":{"total":467,"active_models":357,"scored":140},"snapshot":{"id":347,"created_at":"2026-10-26T00:20:50.593315+00:00"},"attribution":"Arena leaderboard data (lmarena-ai/leaderboard-dataset), CC BY 4.0. Model names mapped to OpenRouter routes by 斩杀线; mapping notes are published per model.","price_source":"models.dev","currency":"USD","price_unit":"per_million_tokens","cost_display":"per_1000_tasks","external_notifications":false}}