{"date":"2026-10-12","description":"340 large language models from GPT-2 to today: organization, release date, open or closed weights, overall index and results on 16 benchmarks — filter, sort and compare side by side.","json":"https://abxk.ai/en/llm/index.json","kind":"section","language":"en","llm_models":{"license":"CC BY 4.0","models":[{"access":"API access","benchmarks":3,"country":"United States of America","key":"claude-haiku-5-5","name":"Claude Haiku 5.5","org":"Anthropic","org_first":"Anthropic","released":"2026-10-07","scores":{"frontiermath":75.1,"gpqa-diamond":89.6,"simpleqa-verified":23.8},"weights":"closed"},{"access":"","benchmarks":1,"country":"France","key":"mistral-large-4","name":"Mistral Large 4","org":"Mistral AI","org_first":"Mistral AI","released":"2026-10-06","scores":{"simpleqa-verified":20},"weights":"unknown"},{"access":"API access","benchmarks":5,"country":"United States of America","eci":166.1,"eci_high":170.6,"eci_low":162.9,"eci_rank":3,"key":"gpt-6-1-sol","name":"GPT-6.1 Sol","org":"OpenAI","org_first":"OpenAI","released":"2026-09-29","scores":{"arc-agi":98.5,"arc-agi-2":94.2,"frontiermath":93.7,"gpqa-diamond":95.4,"simpleqa-verified":73.9},"weights":"closed"},{"access":"API access","benchmarks":3,"country":"United States of America","eci":165,"eci_high":169.1,"eci_low":161.7,"eci_rank":4,"key":"claude-sonnet-5-5","name":"Claude Sonnet 5.5","org":"Anthropic","org_first":"Anthropic","released":"2026-09-28","scores":{"frontiermath":88.8,"gpqa-diamond":95.6,"simpleqa-verified":46.5},"weights":"closed"},{"access":"API access","benchmarks":5,"country":"United States of America","eci":162.7,"eci_high":166.6,"eci_low":160.1,"eci_rank":7,"key":"gpt-6-sol","name":"GPT-6 Sol","org":"OpenAI","org_first":"OpenAI","released":"2026-09-22","scores":{"arc-agi":95.5,"arc-agi-2":89.6,"frontiermath":89.8,"gpqa-diamond":94.3,"simpleqa-verified":60.7},"weights":"closed"},{"access":"API access","benchmarks":5,"country":"United States of America","eci":156.3,"eci_high":158.8,"eci_low":153.7,"eci_rank":24,"key":"gpt-6-luna","name":"GPT-6 Luna","org":"OpenAI","org_first":"OpenAI","released":"2026-09-22","scores":{"arc-agi":86.7,"arc-agi-2":59.3,"frontiermath":78.9,"gpqa-diamond":90.5,"simpleqa-verified":41.4},"weights":"closed"},{"access":"API access","benchmarks":5,"country":"United States of America","eci":167.3,"eci_high":171.7,"eci_low":164.1,"eci_rank":1,"key":"claude-opus-5-5","name":"Claude Opus 5.5","org":"Anthropic","org_first":"Anthropic","released":"2026-09-22","scores":{"arc-agi":98.5,"arc-agi-2":93.3,"frontiermath":91.2,"gpqa-diamond":90.6,"simpleqa-verified":72.2},"weights":"closed"},{"access":"API access","benchmarks":3,"country":"United States of America","eci":153.5,"eci_high":155.8,"eci_low":151.8,"eci_rank":41,"key":"grok-4-7","name":"Grok 4.7","org":"xAI","org_first":"xAI","released":"2026-09-21","scores":{"frontiermath":53,"gpqa-diamond":92.7,"simpleqa-verified":56},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":2,"country":"China","eci":154.9,"eci_high":157.3,"eci_low":148.5,"eci_rank":33,"key":"deepseek-v4-1-flash","name":"DeepSeek V4.1 Flash","org":"DeepSeek","org_first":"DeepSeek","released":"2026-09-09","scores":{"frontiermath":67.4,"gpqa-diamond":89.8},"weights":"open"},{"access":"API access","benchmarks":6,"country":"United States of America","eci":166.4,"eci_high":171.5,"eci_low":163.3,"eci_rank":2,"key":"gpt-6-astra","name":"GPT-6 Astra","org":"OpenAI","org_first":"OpenAI","released":"2026-09-03","scores":{"arc-agi":98.5,"arc-agi-2":95,"frontiermath":93.7,"gpqa-diamond":95.8,"hle":54.8,"simpleqa-verified":75.6},"weights":"closed"},{"access":"API access","benchmarks":1,"country":"United States of America","eci":156.8,"eci_high":159.3,"eci_low":154.8,"eci_rank":19,"key":"muse-spark-1-3","name":"Muse Spark 1.3","org":"Meta AI","org_first":"Meta AI","released":"2026-09-02","scores":{"frontiermath":74.4},"weights":"closed"},{"access":"API access","benchmarks":6,"country":"United States of America","eci":156.7,"eci_high":160.3,"eci_low":154.4,"eci_rank":20,"key":"gemini-3-8-flash","name":"Gemini 3.8 Flash","org":"Google DeepMind","org_first":"Google DeepMind","released":"2026-09-02","scores":{"arc-agi":98.5,"arc-agi-2":89.2,"frontiermath":68.4,"gpqa-diamond":95.4,"hle":44.5,"simpleqa-verified":69.7},"weights":"closed"},{"access":"API access","benchmarks":3,"country":"China","eci":155.1,"eci_high":157.2,"eci_low":153.3,"eci_rank":31,"key":"qwen3-8-max-0902","name":"Qwen3.8 Max (0902)","org":"Alibaba","org_first":"Alibaba","released":"2026-09-01","scores":{"frontiermath":65.6,"gpqa-diamond":92.3,"simpleqa-verified":47.3},"weights":"closed"},{"access":"API access","benchmarks":5,"country":"United States of America","eci":164.7,"eci_high":168.8,"eci_low":161.6,"eci_rank":5,"key":"claude-fable-5-1","name":"Claude Fable 5.1","org":"Anthropic","org_first":"Anthropic","released":"2026-09-01","scores":{"arc-agi":97.5,"arc-agi-2":90,"frontiermath":90.2,"hle":46.5,"simpleqa-verified":70.8},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":4,"country":"China","eci":151.9,"eci_high":153.9,"eci_low":149.6,"eci_rank":47,"key":"glm-5-3-flash","name":"GLM-5.3-Flash","org":"Z.ai (Zhipu AI)","org_first":"Z.ai (Zhipu AI)","released":"2026-08-20","scores":{"arc-agi":91,"arc-agi-2":65.8,"frontiermath":55.8,"gpqa-diamond":90.2},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":2,"country":"China","eci":149.4,"eci_high":151.1,"eci_low":147.3,"eci_rank":58,"key":"qwen-3-8-27b","name":"Qwen 3.8 27B","org":"Alibaba","org_first":"Alibaba","released":"2026-08-14","scores":{"arc-agi":87.5,"arc-agi-2":42.4},"weights":"open"},{"access":"API access","benchmarks":3,"country":"China","eci":155.6,"eci_high":157.9,"eci_low":153.6,"eci_rank":27,"key":"glm-5-3","name":"GLM-5.3","org":"Z.ai (Zhipu AI)","org_first":"Z.ai (Zhipu AI)","released":"2026-08-14","scores":{"frontiermath":68.8,"gpqa-diamond":90.9,"simpleqa-verified":41},"weights":"closed"},{"access":"API access","benchmarks":5,"country":"United States of America","eci":157.3,"eci_high":159.9,"eci_low":155.4,"eci_rank":16,"key":"gemini-3-7-flash","name":"Gemini 3.7 Flash","org":"Google DeepMind","org_first":"Google DeepMind","released":"2026-08-13","scores":{"arc-agi":95.5,"arc-agi-2":84.6,"frontiermath":71.6,"gpqa-diamond":94.8,"simpleqa-verified":69.2},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":5,"country":"China","eci":155.3,"eci_high":157.4,"eci_low":153.5,"eci_rank":29,"key":"deepseek-v4-pro-0813","name":"DeepSeek V4 Pro 0813","org":"DeepSeek","org_first":"DeepSeek","released":"2026-08-13","scores":{"arc-agi":90.5,"arc-agi-2":61.3,"frontiermath":64.6,"gpqa-diamond":91.7,"simpleqa-verified":52.9},"weights":"open"},{"access":"API access","benchmarks":5,"country":"United States of America","eci":156.4,"eci_high":159,"eci_low":154.5,"eci_rank":23,"key":"grok-4-6","name":"Grok 4.6","org":"xAI","org_first":"xAI","released":"2026-08-12","scores":{"arc-agi":87.5,"arc-agi-2":67.1,"frontiermath":66,"gpqa-diamond":94,"simpleqa-verified":49.3},"weights":"closed"},{"access":"API access","benchmarks":1,"country":"United States of America","eci":154.9,"eci_high":157.4,"eci_low":152.8,"eci_rank":32,"key":"muse-spark-1-2","name":"Muse Spark 1.2","org":"Meta AI","org_first":"Meta AI","released":"2026-08-05","scores":{"simpleqa-verified":60.3},"weights":"closed"},{"access":"API access","benchmarks":3,"country":"China","eci":156.4,"eci_high":158.6,"eci_low":154.3,"eci_rank":22,"key":"qwen-3-8-max","name":"Qwen 3.8 Max","org":"Alibaba","org_first":"Alibaba","released":"2026-08-02","scores":{"frontiermath":74.7,"gpqa-diamond":92.7,"simpleqa-verified":45.8},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":5,"country":"China","eci":154.5,"eci_high":156.3,"eci_low":152.1,"eci_rank":36,"key":"deepseek-v4-flash-0731","name":"DeepSeek V4 Flash 0731","org":"DeepSeek","org_first":"DeepSeek","released":"2026-07-31","scores":{"arc-agi":89,"arc-agi-2":61.4,"frontiermath":57.5,"gpqa-diamond":91,"simpleqa-verified":33.6},"weights":"open"},{"access":"API access","benchmarks":2,"country":"China","eci":144.6,"eci_high":147.3,"eci_low":142.5,"eci_rank":88,"key":"qwen3-7-flash","name":"Qwen3.7 Flash","org":"Alibaba","org_first":"Alibaba","released":"2026-07-27","scores":{"frontiermath":19.3,"gpqa-diamond":82.3},"weights":"closed"},{"access":"API access","benchmarks":5,"country":"United States of America","eci":162.8,"eci_high":166.5,"eci_low":160.1,"eci_rank":6,"key":"claude-opus-5","name":"Claude Opus 5","org":"Anthropic","org_first":"Anthropic","released":"2026-07-24","scores":{"arc-agi":97.5,"arc-agi-2":90.4,"frontiermath":85.6,"gpqa-diamond":93.9,"simpleqa-verified":59.9},"weights":"closed"},{"access":"API access","benchmarks":5,"country":"United States of America","eci":154.2,"eci_high":156.2,"eci_low":152.6,"eci_rank":38,"key":"gemini-3-6-flash","name":"Gemini 3.6 Flash","org":"Google DeepMind","org_first":"Google DeepMind","released":"2026-07-21","scores":{"arc-agi":91.2,"arc-agi-2":60.4,"frontiermath":58.9,"gpqa-diamond":94.1,"simpleqa-verified":66.2},"weights":"closed"},{"access":"API access","benchmarks":4,"country":"United States of America","eci":145.1,"eci_high":146.6,"eci_low":142.9,"eci_rank":86,"key":"gemini-3-5-flash-lite","name":"Gemini 3.5 Flash-Lite","org":"Google DeepMind","org_first":"Google DeepMind","released":"2026-07-21","scores":{"arc-agi":53.5,"arc-agi-2":10.3,"frontiermath":26,"gpqa-diamond":83.3},"weights":"closed"},{"access":"Open weights (non-commercial)","benchmarks":5,"country":"China","eci":157.4,"eci_high":160,"eci_low":155.2,"eci_rank":15,"key":"kimi-k3","name":"Kimi K3","org":"Moonshot","org_first":"Moonshot","released":"2026-07-16","scores":{"arc-agi":94.5,"arc-agi-2":60.4,"frontiermath":72.2,"gpqa-diamond":93.1,"simpleqa-verified":50.6},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":5,"country":"United States of America","eci":150.2,"eci_high":152,"eci_low":147.4,"eci_rank":52,"key":"inkling-small","name":"Inkling-Small","org":"Thinking Machines","org_first":"Thinking Machines","released":"2026-07-15","scores":{"arc-agi":84,"arc-agi-2":40.1,"frontiermath":46.3,"gpqa-diamond":88.5,"simpleqa-verified":19.1},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":5,"country":"United States of America","eci":148.5,"eci_high":150.4,"eci_low":145.7,"eci_rank":63,"key":"inkling","name":"Inkling","org":"Thinking Machines","org_first":"Thinking Machines","released":"2026-07-15","scores":{"arc-agi":79.5,"arc-agi-2":36.5,"frontiermath":33.3,"gpqa-diamond":88.3,"simpleqa-verified":40.3},"weights":"open"},{"access":"API access","benchmarks":1,"country":"United States of America","eci":154.2,"eci_high":156.8,"eci_low":152,"eci_rank":37,"key":"muse-spark-1-1","name":"Muse Spark 1.1","org":"Meta AI","org_first":"Meta AI","released":"2026-07-09","scores":{"simpleqa-verified":57.8},"weights":"closed"},{"access":"API access","benchmarks":5,"country":"United States of America","eci":159.6,"eci_high":162.7,"eci_low":157.2,"eci_rank":11,"key":"gpt-5-6-terra","name":"GPT-5.6 Terra","org":"OpenAI","org_first":"OpenAI","released":"2026-07-09","scores":{"arc-agi":96.5,"arc-agi-2":83.9,"frontiermath":86,"gpqa-diamond":93.3,"simpleqa-verified":43.2},"weights":"closed"},{"access":"API access","benchmarks":5,"country":"United States of America","eci":161.7,"eci_high":165.1,"eci_low":159.3,"eci_rank":10,"key":"gpt-5-6-sol","name":"GPT-5.6 Sol","org":"OpenAI","org_first":"OpenAI","released":"2026-07-09","scores":{"arc-agi":97.5,"arc-agi-2":92.5,"frontiermath":89.1,"gpqa-diamond":93.5,"simpleqa-verified":69.7},"weights":"closed"},{"access":"API access","benchmarks":5,"country":"United States of America","eci":156.4,"eci_high":158.7,"eci_low":154,"eci_rank":21,"key":"gpt-5-6-luna","name":"GPT-5.6 Luna","org":"OpenAI","org_first":"OpenAI","released":"2026-07-09","scores":{"arc-agi":88,"arc-agi-2":59.5,"frontiermath":82.1,"gpqa-diamond":91.6,"simpleqa-verified":41},"weights":"closed"},{"access":"API access","benchmarks":5,"country":"United States of America","eci":153.9,"eci_high":155.6,"eci_low":152.2,"eci_rank":39,"key":"grok-4-5","name":"Grok 4.5","org":"xAI","org_first":"xAI","released":"2026-07-08","scores":{"arc-agi":87.2,"arc-agi-2":52.6,"frontiermath":57.2,"gpqa-diamond":93.4,"simpleqa-verified":48.3},"weights":"closed"},{"access":"API access","benchmarks":3,"country":"United States of America","eci":156.2,"eci_high":158.8,"eci_low":153.5,"eci_rank":26,"key":"claude-sonnet-5","name":"Claude Sonnet 5","org":"Anthropic","org_first":"Anthropic","released":"2026-06-30","scores":{"frontiermath":65.6,"gpqa-diamond":90.5,"simpleqa-verified":33.7},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":6,"country":"China","eci":151.8,"eci_high":153.8,"eci_low":149.8,"eci_rank":49,"key":"glm-5-2","name":"GLM-5.2","org":"Z.ai (Zhipu AI)","org_first":"Z.ai (Zhipu AI)","released":"2026-06-16","scores":{"arc-agi":77,"arc-agi-2":22.8,"frontiermath":59.2,"gpqa-diamond":91.9,"simpleqa-verified":34.2,"swe-bench-verified":78.7},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":3,"country":"China","eci":150,"eci_high":151.5,"eci_low":147.9,"eci_rank":55,"key":"kimi-k2-7-code","name":"Kimi K2.7 Code","org":"Moonshot","org_first":"Moonshot","released":"2026-06-12","scores":{"frontiermath":54,"gpqa-diamond":87.9,"simpleqa-verified":36.5},"weights":"open"},{"access":"API access","benchmarks":5,"country":"United States of America","eci":162.1,"eci_high":166.2,"eci_low":159.3,"eci_rank":9,"key":"claude-fable-5","name":"Claude Fable 5","org":"Anthropic","org_first":"Anthropic","released":"2026-06-09","scores":{"arc-agi":98.5,"arc-agi-2":89.2,"frontiermath":87,"gpqa-diamond":85.9,"simpleqa-verified":70.7},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":1,"country":"United States of America","eci":146.2,"eci_high":148,"eci_low":143.9,"eci_rank":77,"key":"nemotron-3-ultra","name":"Nemotron 3 Ultra","org":"Nvidia","org_first":"Nvidia","released":"2026-06-04","scores":{"gpqa-diamond":85.4},"weights":"open"},{"access":"API access","benchmarks":2,"country":"China","eci":147.4,"eci_high":148.9,"eci_low":145.8,"eci_rank":67,"key":"qwen3-7-plus","name":"Qwen3.7-Plus","org":"Alibaba","org_first":"Alibaba","released":"2026-06-02","scores":{"frontiermath":34.4,"gpqa-diamond":87.9},"weights":"closed"},{"access":"Open weights (restricted use)","benchmarks":1,"country":"China","eci":146.9,"eci_high":149.4,"eci_low":142.6,"eci_rank":69,"key":"minimax-m3","name":"MiniMax-M3","org":"MiniMax","org_first":"MiniMax","released":"2026-06-01","scores":{"gpqa-diamond":90.9},"weights":"open"},{"access":"API access","benchmarks":5,"country":"United States of America","eci":158.2,"eci_high":160.8,"eci_low":156,"eci_rank":14,"key":"claude-opus-4-8","name":"Claude Opus 4.8","org":"Anthropic","org_first":"Anthropic","released":"2026-05-28","scores":{"arc-agi":92.5,"arc-agi-2":72.1,"frontiermath":80,"gpqa-diamond":91,"simpleqa-verified":53},"weights":"closed"},{"access":"API access","benchmarks":4,"country":"China","eci":153.7,"eci_high":155.8,"eci_low":151.7,"eci_rank":40,"key":"qwen3-7-max","name":"Qwen3.7-Max","org":"Alibaba","org_first":"Alibaba","released":"2026-05-19","scores":{"frontiermath":64.6,"gpqa-diamond":90.9,"simpleqa-verified":55.8,"swe-bench-verified":77.3},"weights":"closed"},{"access":"API access","benchmarks":6,"country":"United States of America","eci":154.5,"eci_high":156.5,"eci_low":152.5,"eci_rank":35,"key":"gemini-3-5-flash","name":"Gemini 3.5 Flash","org":"Google DeepMind","org_first":"Google DeepMind","released":"2026-05-19","scores":{"arc-agi":92.5,"arc-agi-2":72.1,"frontiermath":62.8,"gpqa-diamond":92.8,"simpleqa-verified":66.2,"swe-bench-verified":79.3},"weights":"closed"},{"access":"API access","benchmarks":2,"country":"United States of America","eci":142.5,"eci_high":145.3,"eci_low":139.8,"eci_rank":103,"key":"gpt-5-5-instant","name":"GPT-5.5 Instant","org":"OpenAI","org_first":"OpenAI","released":"2026-05-05","scores":{"frontiermath":26.3,"gpqa-diamond":82.5},"weights":"closed"},{"access":"Open weights (restricted use)","benchmarks":0,"country":"France","eci":141.3,"eci_high":143.2,"eci_low":138.6,"eci_rank":111,"key":"mistral-medium-3-5","name":"Mistral Medium 3.5","org":"Mistral AI","org_first":"Mistral AI","released":"2026-04-28","scores":{},"weights":"open"},{"access":"Unknown","benchmarks":3,"country":"China","eci":143.3,"eci_high":144.9,"eci_low":140.6,"eci_rank":97,"key":"qwen-3-6-flash","name":"Qwen 3.6 Flash","org":"Alibaba","org_first":"Alibaba","released":"2026-04-27","scores":{"frontiermath":22.5,"gpqa-diamond":83.3,"simpleqa-verified":15.9},"weights":"unknown"},{"access":"Open weights (unrestricted)","benchmarks":4,"country":"China","eci":149.1,"eci_high":150.7,"eci_low":147.3,"eci_rank":61,"key":"deepseek-v4-pro","name":"DeepSeek-V4-Pro","org":"DeepSeek","org_first":"DeepSeek","released":"2026-04-24","scores":{"frontiermath":45.3,"gpqa-diamond":90.9,"simpleqa-verified":47,"swe-bench-verified":77.6},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":0,"country":"China","eci":146.1,"eci_high":147.8,"eci_low":143.9,"eci_rank":78,"key":"deepseek-v4-flash","name":"DeepSeek-V4-Flash","org":"DeepSeek","org_first":"DeepSeek","released":"2026-04-24","scores":{},"weights":"open"},{"access":"API access","benchmarks":4,"country":"United States of America","eci":162.1,"eci_high":165.8,"eci_low":159,"eci_rank":8,"key":"gpt-5-5-pro","name":"GPT-5.5 Pro","org":"OpenAI","org_first":"OpenAI","released":"2026-04-23","scores":{"arc-agi":96.5,"arc-agi-2":84.6,"frontiermath":87.7,"gpqa-diamond":93.9},"weights":"closed"},{"access":"API access","benchmarks":7,"country":"United States of America","eci":159.1,"eci_high":162.3,"eci_low":156.7,"eci_rank":12,"key":"gpt-5-5","name":"GPT-5.5","org":"OpenAI","org_first":"OpenAI","released":"2026-04-23","scores":{"arc-agi":95,"arc-agi-2":85,"frontiermath":85.3,"gpqa-diamond":94,"simpleqa-verified":63,"swe-bench-verified":80.6,"terminal-bench":84.7},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":2,"country":"China","eci":146.5,"eci_high":147.7,"eci_low":144.3,"eci_rank":74,"key":"qwen3-6-27b","name":"Qwen3.6 27B","org":"Alibaba","org_first":"Alibaba","released":"2026-04-22","scores":{"frontiermath":35.1,"gpqa-diamond":85.9},"weights":"open"},{"access":"Unknown","benchmarks":3,"country":"China","eci":149.2,"eci_high":152.1,"eci_low":147.4,"eci_rank":60,"key":"qwen-3-6-max-preview","name":"Qwen 3.6 Max (Preview)","org":"Alibaba","org_first":"Alibaba","released":"2026-04-20","scores":{"gpqa-diamond":87.4,"simpleqa-verified":52,"swe-bench-verified":76.7},"weights":"unknown"},{"access":"Open weights (unrestricted)","benchmarks":4,"country":"China","eci":151.1,"eci_high":152.8,"eci_low":149.3,"eci_rank":50,"key":"kimi-k2-6","name":"Kimi K2.6","org":"Moonshot","org_first":"Moonshot","released":"2026-04-20","scores":{"frontiermath":57.2,"gpqa-diamond":90.8,"simpleqa-verified":34.9,"swe-bench-verified":76.7},"weights":"open"},{"access":"Hosted access (no API)","benchmarks":3,"country":"United States of America","eci":149.2,"eci_high":150.7,"eci_low":147.6,"eci_rank":59,"key":"grok-4-3-beta","name":"Grok 4.3 Beta","org":"xAI","org_first":"xAI","released":"2026-04-17","scores":{"frontiermath":42.8,"gpqa-diamond":88.8,"simpleqa-verified":33.2},"weights":"closed"},{"access":"API access","benchmarks":8,"country":"United States of America","eci":156.2,"eci_high":158.4,"eci_low":154.4,"eci_rank":25,"key":"claude-opus-4-7","name":"Claude Opus 4.7","org":"Anthropic","org_first":"Anthropic","released":"2026-04-16","scores":{"arc-agi":93.5,"arc-agi-2":75.8,"frontiermath":70.2,"gpqa-diamond":90.2,"hle":36.2,"simpleqa-verified":51.7,"swe-bench-verified":83.5,"terminal-bench":80.2},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":3,"country":"China","eci":143.9,"eci_high":145.9,"eci_low":141.2,"eci_rank":94,"key":"qwen-3-6-35b-a3b","name":"Qwen 3.6 35B-A3B","org":"Alibaba","org_first":"Alibaba","released":"2026-04-14","scores":{"frontiermath":20.4,"gpqa-diamond":84.8,"terminal-bench":23},"weights":"open"},{"access":"API access","benchmarks":2,"country":"United States of America","eci":152,"eci_high":155.5,"eci_low":150.1,"eci_rank":46,"key":"muse-spark","name":"Muse Spark","org":"Meta AI","org_first":"Meta AI","released":"2026-04-08","scores":{"gpqa-diamond":89.8,"hle":40.6},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":4,"country":"China","eci":149.8,"eci_high":151.4,"eci_low":147.9,"eci_rank":56,"key":"glm-5-1","name":"GLM-5.1","org":"Z.ai (Zhipu AI)","org_first":"Z.ai (Zhipu AI)","released":"2026-04-07","scores":{"frontiermath":36.8,"gpqa-diamond":89.9,"simpleqa-verified":34,"swe-bench-verified":74.2},"weights":"open"},{"access":"Limited access","benchmarks":1,"country":"United States of America","key":"claude-mythos-preview","name":"Claude Mythos Preview","org":"Anthropic","org_first":"Anthropic","released":"2026-04-07","scores":{"metr-time-horizon":1044.78},"weights":"closed"},{"access":"Open weights (restricted use)","benchmarks":2,"country":"United States of America","eci":142.7,"eci_high":145.1,"eci_low":140.2,"eci_rank":100,"key":"gemma-4-31b-it","name":"Gemma 4 31B IT","org":"Google DeepMind","org_first":"Google DeepMind","released":"2026-04-02","scores":{"gpqa-diamond":75.8,"simpleqa-verified":10.4},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":1,"country":"United States of America","eci":141.8,"eci_high":143.4,"eci_low":138.7,"eci_rank":107,"key":"gemma-4-26b-a4b","name":"Gemma 4 26B A4B","org":"Google DeepMind","org_first":"Google DeepMind","released":"2026-04-02","scores":{"gpqa-diamond":73.2},"weights":"open"},{"access":"API access","benchmarks":4,"country":"China","eci":147.7,"eci_high":149.4,"eci_low":145.7,"eci_rank":65,"key":"qwen-3-6-plus","name":"Qwen 3.6 Plus","org":"Alibaba","org_first":"Alibaba","released":"2026-03-31","scores":{"frontiermath":38.2,"gpqa-diamond":88.4,"simpleqa-verified":44.1,"swe-bench-verified":57.9},"weights":"closed"},{"access":"Open weights (non-commercial)","benchmarks":1,"country":"China","eci":145.8,"eci_high":147.9,"eci_low":137.9,"eci_rank":82,"key":"minimax-m2-7","name":"MiniMax-M2.7","org":"MiniMax","org_first":"MiniMax","released":"2026-03-18","scores":{"terminal-bench":45.1},"weights":"open"},{"access":"API access","benchmarks":5,"country":"United States of America","eci":145.8,"eci_high":147.6,"eci_low":143.5,"eci_rank":81,"key":"gpt-5-4-nano","name":"GPT-5.4 Nano","org":"OpenAI","org_first":"OpenAI","released":"2026-03-17","scores":{"arc-agi":51.5,"arc-agi-2":5.7,"frontiermath":44.9,"gpqa-diamond":78.5,"simpleqa-verified":11.7},"weights":"closed"},{"access":"API access","benchmarks":5,"country":"United States of America","eci":148.8,"eci_high":150.3,"eci_low":147.1,"eci_rank":62,"key":"gpt-5-4-mini","name":"GPT-5.4 Mini","org":"OpenAI","org_first":"OpenAI","released":"2026-03-17","scores":{"arc-agi":63.7,"arc-agi-2":18.9,"frontiermath":51.2,"gpqa-diamond":86.9,"simpleqa-verified":29.4},"weights":"closed"},{"access":"API access","benchmarks":6,"country":"United States of America","eci":158.9,"eci_high":161.8,"eci_low":156.6,"eci_rank":13,"key":"gpt-5-4-pro","name":"GPT-5.4 Pro","org":"OpenAI","org_first":"OpenAI","released":"2026-03-05","scores":{"arc-agi":94.5,"arc-agi-2":83.3,"frontiermath":82.5,"gpqa-diamond":94.6,"hle":44.3,"simpleqa-verified":46.3},"weights":"closed"},{"access":"API access","benchmarks":9,"country":"United States of America","eci":156.8,"eci_high":158.9,"eci_low":155,"eci_rank":18,"key":"gpt-5-4","name":"GPT-5.4","org":"OpenAI","org_first":"OpenAI","released":"2026-03-05","scores":{"arc-agi":93.7,"arc-agi-2":74,"frontiermath":78.6,"gpqa-diamond":93.3,"hle":36.2,"metr-time-horizon":341.74,"simpleqa-verified":45.1,"swe-bench-verified":76.9,"terminal-bench":81.8},"weights":"closed"},{"access":"API access","benchmarks":3,"country":"United States of America","eci":144.5,"eci_high":146,"eci_low":142.4,"eci_rank":89,"key":"gemini-3-1-flash-lite","name":"Gemini 3.1 Flash-Lite","org":"Google","org_first":"Google","released":"2026-03-03","scores":{"frontiermath":27.7,"gpqa-diamond":81.8,"hle":8.6},"weights":"closed"},{"access":"Unknown","benchmarks":3,"country":"China","eci":144,"eci_high":145.7,"eci_low":141.8,"eci_rank":93,"key":"qwen-3-5-flash-hosted-35b-a3b","name":"Qwen 3.5 Flash (hosted 35B-A3B)","org":"Alibaba","org_first":"Alibaba","released":"2026-02-25","scores":{"frontiermath":18.2,"gpqa-diamond":82.3,"simpleqa-verified":20.3},"weights":"unknown"},{"access":"Open weights (restricted use)","benchmarks":2,"country":"China","eci":139.5,"eci_high":141.5,"eci_low":136.2,"eci_rank":121,"key":"qwen3-5-9b","name":"Qwen3.5-9B","org":"Alibaba","org_first":"Alibaba","released":"2026-02-24","scores":{"gpqa-diamond":79,"terminal-bench":9.2},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":1,"country":"China","eci":142.5,"eci_high":144.8,"eci_low":140.3,"eci_rank":102,"key":"qwen3-5-35b-a3b","name":"Qwen3.5-35B-A3B","org":"Alibaba","org_first":"Alibaba","released":"2026-02-24","scores":{"gpqa-diamond":83.5},"weights":"open"},{"access":"API access","benchmarks":9,"country":"United States of America","eci":154.8,"eci_high":157.3,"eci_low":152.4,"eci_rank":34,"key":"gemini-3-1-pro","name":"Gemini 3.1 Pro","org":"Google DeepMind","org_first":"Google DeepMind","released":"2026-02-19","scores":{"arc-agi":98,"arc-agi-2":77.1,"frontiermath":59.6,"gpqa-diamond":94.4,"hle":46.4,"metr-time-horizon":384.15,"simpleqa-verified":73.5,"swe-bench-verified":75.6,"terminal-bench":80.2},"weights":"closed"},{"access":"API access","benchmarks":6,"country":"United States of America","eci":152,"eci_high":154.5,"eci_low":149.4,"eci_rank":45,"key":"grok-4-20","name":"Grok 4.20","org":"xAI","org_first":"xAI","released":"2026-02-17","scores":{"arc-agi":89.5,"arc-agi-2":65.1,"frontiermath":44.9,"gpqa-diamond":89.3,"simpleqa-verified":30.2,"terminal-bench":57.3},"weights":"closed"},{"access":"API access","benchmarks":6,"country":"United States of America","eci":152.2,"eci_high":154.5,"eci_low":149.9,"eci_rank":44,"key":"claude-sonnet-4-6","name":"Claude Sonnet 4.6","org":"Anthropic","org_first":"Anthropic","released":"2026-02-17","scores":{"arc-agi":86.5,"arc-agi-2":60.4,"gpqa-diamond":87.4,"simpleqa-verified":35.5,"swe-bench-verified":75.2,"terminal-bench":53.4},"weights":"closed"},{"access":"Unknown","benchmarks":2,"country":"China","eci":146.8,"eci_high":147.9,"eci_low":144.7,"eci_rank":71,"key":"qwen-3-5-plus-hosted-397b-a17b","name":"Qwen 3.5 Plus (hosted 397B-A17B)","org":"Alibaba","org_first":"Alibaba","released":"2026-02-16","scores":{"gpqa-diamond":84.8,"simpleqa-verified":25.4},"weights":"unknown"},{"access":"Open weights (unrestricted)","benchmarks":2,"country":"China","eci":146.7,"eci_high":148.3,"eci_low":144.9,"eci_rank":73,"key":"qwen3-5-397b-a17b","name":"Qwen3.5 397B-A17B","org":"Alibaba","org_first":"Alibaba","released":"2026-02-13","scores":{"frontiermath":31.2,"gpqa-diamond":86.4},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":3,"country":"China","eci":146.7,"eci_high":147.9,"eci_low":141.9,"eci_rank":72,"key":"minimax-m2-5","name":"MiniMax-M2.5","org":"MiniMax","org_first":"MiniMax","released":"2026-02-12","scores":{"arc-agi":63.7,"arc-agi-2":4.9,"terminal-bench":42.7},"weights":"open"},{"access":"Hosted access (no API)","benchmarks":2,"country":"United States of America","key":"gemini-3-deep-think","name":"Gemini 3 Deep Think","org":"Google DeepMind","org_first":"Google DeepMind","released":"2026-02-12","scores":{"arc-agi":96,"arc-agi-2":84.6},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":5,"country":"China","eci":145.8,"eci_high":147.7,"eci_low":143.8,"eci_rank":80,"key":"glm-5","name":"GLM-5","org":"Z.ai (Zhipu AI)","org_first":"Z.ai (Zhipu AI)","released":"2026-02-11","scores":{"arc-agi":44.7,"arc-agi-2":4.9,"gpqa-diamond":87.8,"swe-bench-verified":72.1,"terminal-bench":52.4},"weights":"open"},{"access":"API access","benchmarks":3,"country":"United States of America","eci":156.8,"eci_high":160.7,"eci_low":152.8,"eci_rank":17,"key":"gpt-5-3-codex","name":"GPT-5.3 Codex","org":"OpenAI","org_first":"OpenAI","released":"2026-02-05","scores":{"metr-time-horizon":349.53,"swe-bench-verified":74.8,"terminal-bench":78.4},"weights":"closed"},{"access":"API access","benchmarks":9,"country":"United States of America","eci":155.2,"eci_high":157.3,"eci_low":153.4,"eci_rank":30,"key":"claude-opus-4-6","name":"Claude Opus 4.6","org":"Anthropic","org_first":"Anthropic","released":"2026-02-05","scores":{"arc-agi":94,"arc-agi-2":69.2,"frontiermath":66,"gpqa-diamond":90.5,"hle":34.4,"metr-time-horizon":718.81,"simpleqa-verified":47,"swe-bench-verified":78.7,"terminal-bench":79.8},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":7,"country":"China","eci":148,"eci_high":149.3,"eci_low":146.4,"eci_rank":64,"key":"kimi-k2-5","name":"Kimi K2.5","org":"Moonshot","org_first":"Moonshot","released":"2026-01-27","scores":{"arc-agi":65.3,"arc-agi-2":11.8,"gpqa-diamond":87.6,"hle":24.4,"simpleqa-verified":34.3,"swe-bench-verified":73.8,"terminal-bench":43.2},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":1,"country":"China","key":"minimax-m2-1","name":"MiniMax-M2.1","org":"MiniMax","org_first":"MiniMax","released":"2025-12-23","scores":{"terminal-bench":36.6},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":3,"country":"China","eci":143.5,"eci_high":145.4,"eci_low":141.1,"eci_rank":96,"key":"glm-4-7","name":"GLM-4.7","org":"Z.ai (Zhipu AI)","org_first":"Z.ai (Zhipu AI)","released":"2025-12-22","scores":{"gpqa-diamond":83.3,"simpleqa-verified":32.2,"terminal-bench":33.4},"weights":"open"},{"access":"API access","benchmarks":1,"country":"United States of America","key":"gpt-5-2-codex","name":"GPT-5.2 Codex","org":"OpenAI","org_first":"OpenAI","released":"2025-12-18","scores":{"terminal-bench":66.5},"weights":"closed"},{"access":"API access","benchmarks":7,"country":"United States of America","eci":151.8,"eci_high":153.3,"eci_low":150.3,"eci_rank":48,"key":"gemini-3-flash","name":"Gemini 3 Flash","org":"Google DeepMind","org_first":"Google DeepMind","released":"2025-12-17","scores":{"arc-agi":84.7,"arc-agi-2":33.6,"frontiermath":51.2,"gpqa-diamond":89.4,"simpleqa-verified":66.8,"swe-bench-verified":75.4,"terminal-bench":64.3},"weights":"closed"},{"access":"API access","benchmarks":3,"country":"United States of America","eci":155.4,"eci_high":158.1,"eci_low":153.1,"eci_rank":28,"key":"gpt-5-2-pro","name":"GPT-5.2 Pro","org":"OpenAI","org_first":"OpenAI","released":"2025-12-11","scores":{"arc-agi":90.5,"arc-agi-2":54.2,"frontiermath":74},"weights":"closed"},{"access":"API access","benchmarks":9,"country":"United States of America","eci":153.4,"eci_high":155.4,"eci_low":151.6,"eci_rank":42,"key":"gpt-5-2","name":"GPT-5.2","org":"OpenAI","org_first":"OpenAI","released":"2025-12-11","scores":{"arc-agi":86.2,"arc-agi-2":52.9,"frontiermath":67.4,"gpqa-diamond":91.4,"hle":27.8,"metr-time-horizon":352.25,"simpleqa-verified":37.1,"swe-bench-verified":73.8,"terminal-bench":64.9},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":5,"country":"China","eci":146.3,"eci_high":147.4,"eci_low":144.5,"eci_rank":76,"key":"deepseek-v3-2","name":"DeepSeek-V3.2","org":"DeepSeek","org_first":"DeepSeek","released":"2025-12-01","scores":{"aider-polyglot":74.2,"arc-agi":57,"arc-agi-2":4,"gpqa-diamond":83.4,"terminal-bench":39.6},"weights":"open"},{"access":"API access","benchmarks":9,"country":"United States of America","eci":150.1,"eci_high":152.8,"eci_low":148,"eci_rank":53,"key":"claude-opus-4-5","name":"Claude Opus 4.5","org":"Anthropic","org_first":"Anthropic","released":"2025-11-24","scores":{"arc-agi":80,"arc-agi-2":37.6,"frontiermath":34.4,"gpqa-diamond":86,"hle":25.2,"metr-time-horizon":292.99,"simpleqa-verified":45.7,"swe-bench-verified":76.7,"terminal-bench":63.1},"weights":"closed"},{"access":"API access","benchmarks":2,"country":"United States of America","key":"gpt-5-1-codex-max","name":"GPT-5.1-Codex-Max","org":"OpenAI","org_first":"OpenAI","released":"2025-11-19","scores":{"metr-time-horizon":223.71,"terminal-bench":60.4},"weights":"closed"},{"access":"API access","benchmarks":7,"country":"United States of America","eci":152.9,"eci_high":155.1,"eci_low":150.9,"eci_rank":43,"key":"gemini-3-pro","name":"Gemini 3 Pro","org":"Google DeepMind","org_first":"Google DeepMind","released":"2025-11-18","scores":{"arc-agi":75,"arc-agi-2":31.1,"gpqa-diamond":92.6,"hle":37.5,"metr-time-horizon":224.33,"swe-bench-verified":72.9,"terminal-bench":69.4},"weights":"closed"},{"access":"API access","benchmarks":7,"country":"United States of America","eci":149.6,"eci_high":151.1,"eci_low":148.2,"eci_rank":57,"key":"gpt-5-1","name":"GPT-5.1","org":"OpenAI","org_first":"OpenAI","released":"2025-11-13","scores":{"arc-agi":72.8,"arc-agi-2":17.6,"gpqa-diamond":87.6,"hle":23.7,"simpleqa-verified":48,"swe-bench-verified":68,"terminal-bench":47.6},"weights":"closed"},{"access":"API access","benchmarks":1,"country":"United States of America","key":"gpt-5-1-codex-mini","name":"GPT-5.1-Codex-mini","org":"OpenAI","org_first":"OpenAI","released":"2025-11-12","scores":{"terminal-bench":61.6},"weights":"closed"},{"access":"API access","benchmarks":1,"country":"United States of America","key":"gpt-5-1-codex","name":"GPT-5.1-Codex","org":"OpenAI","org_first":"OpenAI","released":"2025-11-12","scores":{"terminal-bench":57.8},"weights":"closed"},{"access":"Open weights (restricted use)","benchmarks":3,"country":"China","eci":146,"eci_high":147.5,"eci_low":143.4,"eci_rank":79,"key":"kimi-k2-thinking","name":"Kimi K2 Thinking","org":"Moonshot","org_first":"Moonshot","released":"2025-11-06","scores":{"gpqa-diamond":84.2,"metr-time-horizon":54.19,"terminal-bench":35.7},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":1,"country":"China","key":"minimax-m2","name":"MiniMax-M2","org":"MiniMax","org_first":"MiniMax","released":"2025-10-27","scores":{"terminal-bench":30},"weights":"open"},{"access":"API access","benchmarks":6,"country":"United States of America","eci":142.4,"eci_high":144.1,"eci_low":139.4,"eci_rank":105,"key":"claude-haiku-4-5","name":"Claude Haiku 4.5","org":"Anthropic","org_first":"Anthropic","released":"2025-10-15","scores":{"arc-agi":47.7,"arc-agi-2":4,"gpqa-diamond":71.2,"math-level-5":96.4,"simpleqa-verified":13.2,"terminal-bench":35.5},"weights":"closed"},{"access":"API access","benchmarks":4,"country":"United States of America","eci":150.3,"eci_high":152.8,"eci_low":148.6,"eci_rank":51,"key":"gpt-5-pro","name":"GPT-5 Pro","org":"OpenAI","org_first":"OpenAI","released":"2025-10-07","scores":{"arc-agi":70.2,"arc-agi-2":18.3,"frontiermath":55.8,"hle":31.6},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":1,"country":"China","key":"glm-4-6","name":"GLM-4.6","org":"Z.ai (Zhipu AI), Tsinghua University","org_first":"Z.ai (Zhipu AI)","released":"2025-09-30","scores":{"terminal-bench":24.5},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":1,"country":"China","eci":145,"eci_high":147,"eci_low":142.4,"eci_rank":87,"key":"deepseek-v3-2-exp","name":"DeepSeek-V3.2-Exp","org":"DeepSeek","org_first":"DeepSeek","released":"2025-09-29","scores":{"aider-polyglot":74.2},"weights":"open"},{"access":"API access","benchmarks":10,"country":"United States of America","eci":146.8,"eci_high":148.3,"eci_low":145.1,"eci_rank":70,"key":"claude-sonnet-4-5","name":"Claude Sonnet 4.5","org":"Anthropic","org_first":"Anthropic","released":"2025-09-29","scores":{"arc-agi":63.7,"arc-agi-2":13.6,"frontiermath":23.9,"gpqa-diamond":82.3,"hle":13.7,"math-level-5":97.7,"metr-time-horizon":121.95,"simpleqa-verified":30.7,"swe-bench-verified":71.3,"terminal-bench":46.5},"weights":"closed"},{"access":"API access","benchmarks":1,"country":"United States of America","eci":143,"eci_high":146.4,"eci_low":139.4,"eci_rank":98,"key":"gemini-2-5-flash-sep-2025","name":"Gemini 2.5 Flash (Sep 2025)","org":"Google DeepMind","org_first":"Google DeepMind","released":"2025-09-25","scores":{"terminal-bench":17.1},"weights":"closed"},{"access":"API access","benchmarks":4,"country":"China","eci":142.4,"eci_high":144.8,"eci_low":140.1,"eci_rank":104,"key":"qwen3-max","name":"Qwen3-Max","org":"Alibaba","org_first":"Alibaba","released":"2025-09-24","scores":{"frontiermath":18.9,"gpqa-diamond":72.6,"math-level-5":97.1,"simpleqa-verified":48.7},"weights":"closed"},{"access":"API access","benchmarks":2,"country":"United States of America","eci":144.2,"eci_high":146.6,"eci_low":141.7,"eci_rank":91,"key":"grok-4-fast","name":"Grok 4 Fast","org":"xAI","org_first":"xAI","released":"2025-09-19","scores":{"arc-agi":48.5,"arc-agi-2":5.3},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":1,"country":"France","eci":131.4,"eci_high":133.5,"eci_low":126,"eci_rank":155,"key":"magistral-small-1-2","name":"Magistral Small 1.2","org":"Mistral AI","org_first":"Mistral AI","released":"2025-09-18","scores":{"gpqa-diamond":47.6},"weights":"open"},{"access":"API access","benchmarks":1,"country":"United States of America","key":"gpt-5-codex","name":"GPT-5-Codex","org":"OpenAI","org_first":"OpenAI","released":"2025-09-15","scores":{"terminal-bench":44.3},"weights":"closed"},{"access":"Open weights (restricted use)","benchmarks":1,"country":"China","key":"kimi-k2-sep-2025","name":"Kimi K2 (Sep 2025)","org":"Moonshot","org_first":"Moonshot","released":"2025-09-05","scores":{"aider-polyglot":59.1},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":0,"country":"China","eci":139.9,"eci_high":143.2,"eci_low":135.4,"eci_rank":119,"key":"deepseek-v3-1","name":"DeepSeek-V3.1","org":"DeepSeek","org_first":"DeepSeek","released":"2025-08-21","scores":{},"weights":"open"},{"access":"API access","benchmarks":7,"country":"United States of America","eci":139.4,"eci_high":141.4,"eci_low":134.1,"eci_rank":122,"key":"gpt-5-nano","name":"GPT-5 nano","org":"OpenAI","org_first":"OpenAI","released":"2025-08-07","scores":{"arc-agi":20.7,"arc-agi-2":2.6,"frontiermath":20,"gpqa-diamond":69.4,"math-level-5":95.2,"simpleqa-verified":11.7,"terminal-bench":21.8},"weights":"closed"},{"access":"API access","benchmarks":9,"country":"United States of America","eci":145.5,"eci_high":147.2,"eci_low":143.6,"eci_rank":84,"key":"gpt-5-mini","name":"GPT-5 mini","org":"OpenAI","org_first":"OpenAI","released":"2025-08-07","scores":{"arc-agi":54.3,"arc-agi-2":4.4,"frontiermath":46.7,"gpqa-diamond":75,"hle":19.4,"math-level-5":97.8,"simpleqa-verified":21.6,"swe-bench-verified":64.7,"terminal-bench":34.8},"weights":"closed"},{"access":"API access","benchmarks":11,"country":"United States of America","eci":150,"eci_rank":54,"key":"gpt-5","name":"GPT-5","org":"OpenAI","org_first":"OpenAI","released":"2025-08-07","scores":{"aider-polyglot":88,"arc-agi":65.7,"arc-agi-2":9.9,"frontiermath":55.4,"gpqa-diamond":86.2,"hle":25.3,"math-level-5":98.1,"metr-time-horizon":203.01,"simpleqa-verified":50.1,"swe-bench-verified":73.6,"terminal-bench":49.6},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":2,"country":"United States of America","eci":137.8,"eci_high":139.6,"eci_low":132.9,"eci_rank":129,"key":"gpt-oss-20b","name":"gpt-oss-20b","org":"OpenAI","org_first":"OpenAI","released":"2025-08-05","scores":{"gpqa-diamond":60.8,"terminal-bench":3.4},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":4,"country":"United States of America","eci":139.9,"eci_high":142.8,"eci_low":135.2,"eci_rank":118,"key":"gpt-oss-120b","name":"gpt-oss-120b","org":"OpenAI","org_first":"OpenAI","released":"2025-08-05","scores":{"aider-polyglot":41.8,"gpqa-diamond":75.8,"metr-time-horizon":41.98,"terminal-bench":18.7},"weights":"open"},{"access":"API access","benchmarks":6,"country":"United States of America","eci":144.1,"eci_high":146,"eci_low":141.7,"eci_rank":92,"key":"claude-opus-4-1","name":"Claude Opus 4.1","org":"Anthropic","org_first":"Anthropic","released":"2025-08-05","scores":{"frontiermath":12.6,"gpqa-diamond":77.3,"hle":11.5,"metr-time-horizon":113.69,"swe-bench-verified":73.3,"terminal-bench":38},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":1,"country":"China","key":"glm-4-5","name":"GLM-4.5","org":"Z.ai (Zhipu AI), Tsinghua University","org_first":"Z.ai (Zhipu AI)","released":"2025-08-03","scores":{"hle":8.3},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":1,"country":"China","key":"qwen3-coder-480b-a35b","name":"Qwen3-Coder-480B-A35B","org":"Alibaba","org_first":"Alibaba","released":"2025-07-31","scores":{"terminal-bench":27.2},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":1,"country":"China","eci":139.6,"eci_high":141,"eci_low":135.5,"eci_rank":120,"key":"qwen3-30b-a3b-thinking-jul-2025","name":"Qwen3-30B-A3B-Thinking (Jul 2025)","org":"Alibaba","org_first":"Alibaba","released":"2025-07-30","scores":{"gpqa-diamond":70.1},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":1,"country":"China","eci":137.4,"eci_high":139.4,"eci_low":131.8,"eci_rank":132,"key":"qwen3-30b-a3b-instruct-jul-2025","name":"Qwen3-30B-A3B-Instruct (Jul 2025)","org":"Alibaba","org_first":"Alibaba","released":"2025-07-29","scores":{"gpqa-diamond":55.6},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":2,"country":"China","eci":143.8,"eci_high":145.3,"eci_low":141.9,"eci_rank":95,"key":"qwen3-235b-a22b-thinking-jul-2025","name":"Qwen3-235B-A22B-Thinking (Jul 2025)","org":"Alibaba","org_first":"Alibaba","released":"2025-07-25","scores":{"gpqa-diamond":80.1,"simpleqa-verified":40.4},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":3,"country":"China","eci":138.9,"eci_high":140.8,"eci_low":135.7,"eci_rank":125,"key":"qwen3-235b-a22b-instruct-jul-2025","name":"Qwen3-235B-A22B-Instruct (Jul 2025)","org":"Alibaba","org_first":"Alibaba","released":"2025-07-25","scores":{"aider-polyglot":59.6,"arc-agi":11,"arc-agi-2":1.2},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":1,"country":"China","key":"glm-4-5-air","name":"GLM-4.5-Air","org":"Z.ai (Zhipu AI), Tsinghua University","org_first":"Z.ai (Zhipu AI)","released":"2025-07-20","scores":{"hle":8.1},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":2,"country":"China","eci":140.1,"eci_high":142,"eci_low":136.6,"eci_rank":116,"key":"kimi-k2-jul-2025","name":"Kimi K2 (Jul 2025)","org":"Moonshot","org_first":"Moonshot","released":"2025-07-12","scores":{"aider-polyglot":59.1,"terminal-bench":27.8},"weights":"open"},{"access":"API access","benchmarks":6,"country":"United States of America","eci":146.4,"eci_high":148.1,"eci_low":144.7,"eci_rank":75,"key":"grok-4","name":"Grok 4","org":"xAI","org_first":"xAI","released":"2025-07-09","scores":{"aider-polyglot":79.6,"arc-agi":66.7,"arc-agi-2":16,"gpqa-diamond":87,"metr-time-horizon":110.08,"terminal-bench":27.2},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":1,"country":"France","eci":131.7,"eci_high":133.7,"eci_low":126.5,"eci_rank":154,"key":"mistral-small-3-2","name":"Mistral Small 3.2","org":"Mistral AI","org_first":"Mistral AI","released":"2025-06-20","scores":{"gpqa-diamond":49.1},"weights":"open"},{"access":"API access","benchmarks":0,"country":"United States of America","eci":133.9,"eci_high":136.5,"eci_low":130,"eci_rank":147,"key":"gemini-2-5-flash-lite-jun-2025","name":"Gemini 2.5 Flash-Lite (Jun 2025)","org":"Google DeepMind","org_first":"Google DeepMind","released":"2025-06-17","scores":{},"weights":"closed"},{"access":"API access","benchmarks":1,"country":"United States of America","eci":140.8,"eci_high":142.1,"eci_low":138.4,"eci_rank":113,"key":"gemini-2-5-flash-jun-2025","name":"Gemini 2.5 Flash (Jun 2025)","org":"Google DeepMind","org_first":"Google DeepMind","released":"2025-06-17","scores":{"terminal-bench":17.1},"weights":"closed"},{"access":"API access","benchmarks":3,"country":"United States of America","eci":147.4,"eci_high":149.9,"eci_low":145.7,"eci_rank":66,"key":"o3-pro","name":"o3-pro","org":"OpenAI","org_first":"OpenAI","released":"2025-06-10","scores":{"aider-polyglot":84.9,"arc-agi":59.3,"arc-agi-2":4.9},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":3,"country":"France","eci":133.2,"eci_high":134.5,"eci_low":128.8,"eci_rank":149,"key":"magistral-small-1-0","name":"Magistral Small 1.0","org":"Mistral AI","org_first":"Mistral AI","released":"2025-06-10","scores":{"arc-agi":5,"arc-agi-2":0,"gpqa-diamond":56.1},"weights":"open"},{"access":"API access","benchmarks":2,"country":"France","key":"magistral-medium-1-1","name":"Magistral Medium 1.1","org":"Mistral AI","org_first":"Mistral AI","released":"2025-06-10","scores":{"arc-agi":6.1,"arc-agi-2":0},"weights":"closed"},{"access":"API access","benchmarks":9,"country":"United States of America","eci":145.3,"eci_high":146.8,"eci_low":143.8,"eci_rank":85,"key":"gemini-2-5-pro-jun-2025","name":"Gemini 2.5 Pro (Jun 2025)","org":"Google DeepMind","org_first":"Google DeepMind","released":"2025-06-05","scores":{"aider-polyglot":83.1,"arc-agi":41,"arc-agi-2":4.9,"frontiermath":24.6,"gpqa-diamond":85.3,"hle":21.6,"metr-time-horizon":38.73,"swe-bench-verified":57.6,"terminal-bench":32.6},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":6,"country":"China","eci":141.3,"eci_high":142.8,"eci_low":138.6,"eci_rank":110,"key":"deepseek-r1-may-2025","name":"DeepSeek-R1 (May 2025)","org":"DeepSeek","org_first":"DeepSeek","released":"2025-05-28","scores":{"aider-polyglot":71.4,"arc-agi":21.2,"arc-agi-2":1.1,"gpqa-diamond":76.3,"math-level-5":96.6,"metr-time-horizon":31.17},"weights":"open"},{"access":"API access","benchmarks":7,"country":"United States of America","eci":141.7,"eci_high":143,"eci_low":138.9,"eci_rank":108,"key":"claude-sonnet-4","name":"Claude Sonnet 4","org":"Anthropic","org_first":"Anthropic","released":"2025-05-22","scores":{"aider-polyglot":61.3,"arc-agi":40,"arc-agi-2":5.9,"gpqa-diamond":79.2,"hle":7.8,"math-level-5":84.4,"metr-time-horizon":74.94},"weights":"closed"},{"access":"API access","benchmarks":8,"country":"United States of America","eci":142.7,"eci_high":144.3,"eci_low":140.1,"eci_rank":99,"key":"claude-opus-4","name":"Claude Opus 4","org":"Anthropic","org_first":"Anthropic","released":"2025-05-22","scores":{"aider-polyglot":72,"arc-agi":35.7,"arc-agi-2":8.6,"gpqa-diamond":76.3,"hle":10.7,"math-level-5":85,"metr-time-horizon":100.37,"swe-bench-verified":70.7},"weights":"closed"},{"access":"API access","benchmarks":4,"country":"United States of America","eci":141.5,"eci_high":142.9,"eci_low":139.2,"eci_rank":109,"key":"gemini-2-5-flash-may-2025","name":"Gemini 2.5 Flash (May 2025)","org":"Google DeepMind","org_first":"Google DeepMind","released":"2025-05-20","scores":{"aider-polyglot":55.1,"arc-agi":33.3,"arc-agi-2":2.5,"hle":11},"weights":"closed"},{"access":"API access","benchmarks":3,"country":"France","eci":134.1,"eci_high":135.5,"eci_low":130.1,"eci_rank":146,"key":"mistral-medium-3","name":"Mistral Medium 3","org":"Mistral AI","org_first":"Mistral AI","released":"2025-05-07","scores":{"gpqa-diamond":59.5,"hle":4.5,"math-level-5":81.6},"weights":"closed"},{"access":"API access","benchmarks":4,"country":"United States of America","eci":142.5,"eci_high":145.1,"eci_low":139,"eci_rank":101,"key":"gemini-2-5-pro-may-2025","name":"Gemini 2.5 Pro (May 2025)","org":"Google DeepMind","org_first":"Google DeepMind","released":"2025-05-06","scores":{"aider-polyglot":76.9,"gpqa-diamond":66.7,"hle":17.8,"math-level-5":95.9},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":1,"country":"China","key":"qwen3-4b","name":"Qwen3-4B","org":"Alibaba","org_first":"Alibaba","released":"2025-04-29","scores":{"gpqa-diamond":52.3},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":1,"country":"China","key":"qwen3-1-7b","name":"Qwen3-1.7B","org":"Alibaba","org_first":"Alibaba","released":"2025-04-29","scores":{"gpqa-diamond":38},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":1,"country":"China","eci":136.2,"eci_high":138.2,"eci_low":130.1,"eci_rank":136,"key":"qwen3-8b","name":"Qwen3-8B","org":"Alibaba","org_first":"Alibaba","released":"2025-04-28","scores":{"gpqa-diamond":56.8},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":2,"country":"China","eci":138.5,"eci_high":140.3,"eci_low":135.1,"eci_rank":126,"key":"qwen3-32b","name":"Qwen3-32B","org":"Alibaba","org_first":"Alibaba","released":"2025-04-28","scores":{"aider-polyglot":40,"gpqa-diamond":65.7},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":1,"country":"China","eci":136.2,"eci_high":138.6,"eci_low":129.7,"eci_rank":135,"key":"qwen3-30b-a3b","name":"Qwen3-30B-A3B","org":"Alibaba","org_first":"Alibaba","released":"2025-04-28","scores":{"gpqa-diamond":61.7},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":3,"country":"China","eci":139.3,"eci_high":140.8,"eci_low":135.5,"eci_rank":123,"key":"qwen3-235b-a22b","name":"Qwen3-235B-A22B","org":"Alibaba","org_first":"Alibaba","released":"2025-04-28","scores":{"aider-polyglot":59.6,"gpqa-diamond":70.7,"math-level-5":68.9},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":1,"country":"China","eci":138.2,"eci_high":140,"eci_low":133.7,"eci_rank":128,"key":"qwen3-14b","name":"Qwen3-14B","org":"Alibaba","org_first":"Alibaba","released":"2025-04-28","scores":{"gpqa-diamond":63.8},"weights":"open"},{"access":"API access","benchmarks":3,"country":"United States of America","eci":140,"eci_high":142.1,"eci_low":136.1,"eci_rank":117,"key":"gemini-2-5-flash-apr-2025","name":"Gemini 2.5 Flash (Apr 2025)","org":"Google DeepMind","org_first":"Google DeepMind","released":"2025-04-17","scores":{"aider-polyglot":47.1,"arc-agi":32.3,"hle":12.1},"weights":"closed"},{"access":"API access","benchmarks":9,"country":"United States of America","eci":145.6,"eci_high":147.4,"eci_low":143.2,"eci_rank":83,"key":"o4-mini","name":"o4-mini","org":"OpenAI","org_first":"OpenAI","released":"2025-04-16","scores":{"aider-polyglot":72,"arc-agi":58.7,"arc-agi-2":6.1,"frontiermath":36.1,"gpqa-diamond":79.6,"hle":18.1,"math-level-5":97.8,"metr-time-horizon":76.51,"simpleqa-verified":19.6},"weights":"closed"},{"access":"API access","benchmarks":10,"country":"United States of America","eci":146.9,"eci_high":148.7,"eci_low":145,"eci_rank":68,"key":"o3","name":"o3","org":"OpenAI","org_first":"OpenAI","released":"2025-04-16","scores":{"aider-polyglot":81.3,"arc-agi":60.8,"arc-agi-2":6.5,"frontiermath":33.3,"gpqa-diamond":81.8,"hle":20.3,"math-level-5":97.8,"metr-time-horizon":119.73,"simpleqa-verified":49.4,"swe-bench-verified":62.3},"weights":"closed"},{"access":"API access","benchmarks":6,"country":"United States of America","eci":129.6,"eci_high":131.9,"eci_low":123.2,"eci_rank":161,"key":"gpt-4-1-nano","name":"GPT-4.1 nano","org":"OpenAI","org_first":"OpenAI","released":"2025-04-14","scores":{"aider-polyglot":8.9,"arc-agi":0,"arc-agi-2":0,"gpqa-diamond":48.9,"math-level-5":70,"simpleqa-verified":6},"weights":"closed"},{"access":"API access","benchmarks":7,"country":"United States of America","eci":135,"eci_high":136.7,"eci_low":131.3,"eci_rank":142,"key":"gpt-4-1-mini","name":"GPT-4.1 mini","org":"OpenAI","org_first":"OpenAI","released":"2025-04-14","scores":{"aider-polyglot":32.4,"arc-agi":3.5,"arc-agi-2":0,"frontiermath":6.7,"gpqa-diamond":65.8,"math-level-5":87.3,"simpleqa-verified":12.7},"weights":"closed"},{"access":"API access","benchmarks":9,"country":"United States of America","eci":136.8,"eci_high":138.6,"eci_low":133.6,"eci_rank":133,"key":"gpt-4-1","name":"GPT-4.1","org":"OpenAI","org_first":"OpenAI","released":"2025-04-14","scores":{"aider-polyglot":52.4,"arc-agi":5.5,"arc-agi-2":0.4,"frontiermath":6,"gpqa-diamond":66.9,"hle":5.4,"math-level-5":83,"simpleqa-verified":31.1,"swe-bench-verified":48.5},"weights":"closed"},{"access":"API access","benchmarks":5,"country":"United States of America","eci":140.3,"eci_high":142,"eci_low":137.5,"eci_rank":115,"key":"grok-3-mini","name":"Grok-3 mini","org":"xAI","org_first":"xAI","released":"2025-04-09","scores":{"aider-polyglot":49.3,"arc-agi":16.5,"arc-agi-2":0.4,"gpqa-diamond":76.3,"math-level-5":90.9},"weights":"closed"},{"access":"API access","benchmarks":5,"country":"United States of America","eci":138.3,"eci_high":139.8,"eci_low":135.3,"eci_rank":127,"key":"grok-3","name":"Grok 3","org":"xAI","org_first":"xAI","released":"2025-04-09","scores":{"aider-polyglot":53.3,"arc-agi":5.5,"arc-agi-2":0,"gpqa-diamond":75.8,"math-level-5":88.7},"weights":"closed"},{"access":"API access","benchmarks":1,"country":"China","key":"qwq-plus","name":"QWQ-Plus","org":"Alibaba","org_first":"Alibaba","released":"2025-04-08","scores":{"gpqa-diamond":65.4},"weights":"closed"},{"access":"Open weights (restricted use)","benchmarks":4,"country":"United States of America","eci":129.6,"eci_high":131.4,"eci_low":124.3,"eci_rank":160,"key":"llama-4-scout","name":"Llama 4 Scout","org":"Meta AI","org_first":"Meta AI","released":"2025-04-05","scores":{"arc-agi":0.5,"arc-agi-2":0,"gpqa-diamond":51.8,"math-level-5":62.3},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":6,"country":"United States of America","eci":132.2,"eci_high":134.3,"eci_low":127.7,"eci_rank":152,"key":"llama-4-maverick","name":"Llama 4 Maverick","org":"Meta AI","org_first":"Meta AI","released":"2025-04-05","scores":{"aider-polyglot":15.6,"arc-agi":4.4,"arc-agi-2":0,"gpqa-diamond":67,"hle":5.7,"math-level-5":73},"weights":"open"},{"access":"API access","benchmarks":1,"country":"United States of America","key":"gpt-4o-mar-2025","name":"GPT-4o (Mar 2025)","org":"OpenAI","org_first":"OpenAI","released":"2025-03-27","scores":{"aider-polyglot":45.3},"weights":"closed"},{"access":"API access","benchmarks":5,"country":"United States of America","eci":144.2,"eci_high":147,"eci_low":141.9,"eci_rank":90,"key":"gemini-2-5-pro-mar-2025","name":"Gemini 2.5 Pro (Mar 2025)","org":"Google DeepMind","org_first":"Google DeepMind","released":"2025-03-25","scores":{"aider-polyglot":72.9,"arc-agi":33,"gpqa-diamond":83.8,"hle":18.2,"math-level-5":95.6},"weights":"closed"},{"access":"Open weights (restricted use)","benchmarks":4,"country":"China","eci":135.9,"eci_high":137.8,"eci_low":132.1,"eci_rank":137,"key":"deepseek-v3-mar-2025","name":"DeepSeek-V3 (Mar 2025)","org":"DeepSeek","org_first":"DeepSeek","released":"2025-03-24","scores":{"aider-polyglot":55.1,"gpqa-diamond":67.6,"math-level-5":75.5,"metr-time-horizon":23.12},"weights":"open"},{"access":"API access","benchmarks":2,"country":"United States of America","key":"o1-pro","name":"o1-pro","org":"OpenAI","org_first":"OpenAI","released":"2025-03-19","scores":{"arc-agi":23.3,"hle":8.1},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":2,"country":"France","eci":127.5,"eci_high":129.2,"eci_low":122.2,"eci_rank":171,"key":"mistral-small-3-1","name":"Mistral Small 3.1","org":"Mistral AI","org_first":"Mistral AI","released":"2025-03-17","scores":{"gpqa-diamond":47.5,"math-level-5":46.8},"weights":"open"},{"access":"Open weights (non-commercial)","benchmarks":1,"country":"Canada","key":"cohere-command-a","name":"Cohere Command A","org":"Cohere","org_first":"Cohere","released":"2025-03-13","scores":{"aider-polyglot":12},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":1,"country":"United States of America","eci":116,"eci_high":123.8,"eci_low":97.6,"eci_rank":214,"key":"gemma-3-4b","name":"Gemma 3 4B","org":"Google DeepMind","org_first":"Google DeepMind","released":"2025-03-12","scores":{"gpqa-diamond":23.2},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":1,"country":"United States of America","key":"gemma-3-1b","name":"Gemma 3 1B","org":"Google DeepMind","org_first":"Google DeepMind","released":"2025-03-12","scores":{"gpqa-diamond":19.9},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":1,"country":"United States of America","eci":123.5,"eci_high":128.5,"eci_low":116,"eci_rank":186,"key":"gemma-3-12b","name":"Gemma 3 12B","org":"Google DeepMind","org_first":"Google DeepMind","released":"2025-03-12","scores":{"gpqa-diamond":39.5},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":3,"country":"United States of America","eci":130,"eci_high":132.2,"eci_low":124.3,"eci_rank":159,"key":"gemma-3-27b","name":"Gemma 3 27B","org":"Google DeepMind","org_first":"Google DeepMind","released":"2025-03-11","scores":{"aider-polyglot":4.9,"gpqa-diamond":47.7,"math-level-5":74},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":2,"country":"China","eci":137.6,"eci_high":141.6,"eci_low":132.9,"eci_rank":130,"key":"qwq-32b","name":"QwQ-32B","org":"Alibaba","org_first":"Alibaba","released":"2025-03-05","scores":{"aider-polyglot":20.9,"gpqa-diamond":65.3},"weights":"open"},{"access":"API access","benchmarks":6,"country":"United States of America","eci":136.7,"eci_high":138.7,"eci_low":133.4,"eci_rank":134,"key":"gpt-4-5","name":"GPT-4.5","org":"OpenAI","org_first":"OpenAI","released":"2025-02-27","scores":{"aider-polyglot":44.9,"arc-agi":10.3,"arc-agi-2":0.8,"gpqa-diamond":68.7,"hle":5.4,"math-level-5":78.6},"weights":"closed"},{"access":"API access","benchmarks":8,"country":"United States of America","eci":141.2,"eci_high":142.7,"eci_low":138.8,"eci_rank":112,"key":"claude-3-7-sonnet","name":"Claude 3.7 Sonnet","org":"Anthropic","org_first":"Anthropic","released":"2025-02-24","scores":{"aider-polyglot":64.9,"arc-agi":28.6,"arc-agi-2":0.9,"gpqa-diamond":79.7,"hle":8,"math-level-5":91.2,"metr-time-horizon":60.39,"swe-bench-verified":61},"weights":"closed"},{"access":"Hosted access (no API)","benchmarks":3,"country":"United States of America","eci":135.1,"eci_high":136.9,"eci_low":131,"eci_rank":141,"key":"gemini-2-0-pro","name":"Gemini 2.0 Pro","org":"Google DeepMind","org_first":"Google DeepMind","released":"2025-02-05","scores":{"aider-polyglot":35.6,"gpqa-diamond":65.7,"math-level-5":83.5},"weights":"closed"},{"access":"API access","benchmarks":3,"country":"United States of America","eci":134.7,"eci_high":136.5,"eci_low":130.8,"eci_rank":145,"key":"gemini-2-0-flash-feb-2025","name":"Gemini 2.0 Flash (Feb 2025)","org":"Google DeepMind, Google","org_first":"Google DeepMind","released":"2025-02-05","scores":{"arc-agi-2":1.3,"gpqa-diamond":64.1,"math-level-5":82.2},"weights":"closed"},{"access":"API access","benchmarks":7,"country":"United States of America","eci":140.3,"eci_high":141.8,"eci_low":137.7,"eci_rank":114,"key":"o3-mini","name":"o3-mini","org":"OpenAI","org_first":"OpenAI","released":"2025-01-31","scores":{"aider-polyglot":60.4,"arc-agi":34.5,"arc-agi-2":3,"frontiermath":18.6,"gpqa-diamond":77,"math-level-5":96.5,"simpleqa-verified":15.3},"weights":"closed"},{"access":"API access","benchmarks":1,"country":"United States of America","key":"gpt-4o-jan-2025","name":"GPT-4o (Jan 2025)","org":"OpenAI","org_first":"OpenAI","released":"2025-01-29","scores":{"aider-polyglot":27.1},"weights":"closed"},{"access":"API access","benchmarks":3,"country":"China","eci":132.5,"eci_high":135.9,"eci_low":127.8,"eci_rank":150,"key":"qwen2-5-max","name":"Qwen2.5-Max","org":"Alibaba","org_first":"Alibaba","released":"2025-01-25","scores":{"aider-polyglot":21.8,"gpqa-diamond":56.1,"math-level-5":67.2},"weights":"closed"},{"access":"API access","benchmarks":2,"country":"China","key":"qwen-plus-jan-2025","name":"Qwen Plus (Jan 2025)","org":"Alibaba","org_first":"Alibaba","released":"2025-01-25","scores":{"gpqa-diamond":48.1,"math-level-5":65.3},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":2,"country":"France","eci":127.1,"eci_high":128.8,"eci_low":121.4,"eci_rank":175,"key":"mistral-small-3","name":"Mistral Small 3","org":"Mistral AI","org_first":"Mistral AI","released":"2025-01-25","scores":{"gpqa-diamond":47.3,"math-level-5":44.8},"weights":"open"},{"access":"API access","benchmarks":3,"country":"United States of America","eci":135.4,"eci_high":138.1,"eci_low":130.3,"eci_rank":140,"key":"gemini-2-0-flash-thinking-jan-2025","name":"Gemini 2.0 Flash Thinking (Jan 2025)","org":"Google DeepMind, Google","org_first":"Google DeepMind","released":"2025-01-21","scores":{"aider-polyglot":18.2,"gpqa-diamond":57.1,"hle":6.6},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":1,"country":"China","eci":137.4,"eci_high":138.7,"eci_low":126.6,"eci_rank":131,"key":"deepseek-r1-distill-qwen-32b","name":"DeepSeek-R1-Distill-Qwen-32B","org":"DeepSeek","org_first":"DeepSeek","released":"2025-01-20","scores":{"gpqa-diamond":64.1},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":2,"country":"China","eci":135.4,"eci_high":138,"eci_low":127,"eci_rank":139,"key":"deepseek-r1-distill-qwen-14b","name":"DeepSeek-R1-Distill-Qwen-14B","org":"DeepSeek","org_first":"DeepSeek","released":"2025-01-20","scores":{"gpqa-diamond":44.7,"math-level-5":87.1},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":1,"country":"China","key":"deepseek-r1-distill-qwen-1-5b","name":"DeepSeek-R1-Distill-Qwen-1.5B","org":"DeepSeek","org_first":"DeepSeek","released":"2025-01-20","scores":{"gpqa-diamond":33.6},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":2,"country":"China","key":"deepseek-r1-distill-llama-70b","name":"DeepSeek-R1-Distill-Llama-70B","org":"DeepSeek","org_first":"DeepSeek","released":"2025-01-20","scores":{"gpqa-diamond":55.7,"math-level-5":89.9},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":6,"country":"China","eci":139,"eci_high":140.3,"eci_low":136.3,"eci_rank":124,"key":"deepseek-r1","name":"DeepSeek-R1","org":"DeepSeek","org_first":"DeepSeek","released":"2025-01-20","scores":{"aider-polyglot":56.9,"arc-agi":15.8,"arc-agi-2":1.3,"gpqa-diamond":71.7,"math-level-5":93.1,"metr-time-horizon":26.93},"weights":"open"},{"access":"Open weights (non-commercial)","benchmarks":1,"country":"France","key":"codestral","name":"Codestral","org":"Mistral AI","org_first":"Mistral AI","released":"2025-01-13","scores":{"aider-polyglot":11.1},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":1,"country":"China,United States of America","key":"eurus-2-7b-prime","name":"Eurus-2-7B-PRIME","org":"Tsinghua University, University of Illinois Urbana-Champaign (UIUC), Shanghai AI Lab, Peking University, Shanghai Jiao Tong University, CUHK Shenzhen Research Institute","org_first":"Tsinghua University","released":"2024-12-31","scores":{"gpqa-diamond":33.9},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":7,"country":"China","eci":132.3,"eci_high":135.7,"eci_low":127.5,"eci_rank":151,"key":"deepseek-v3","name":"DeepSeek-V3","org":"DeepSeek","org_first":"DeepSeek","released":"2024-12-26","scores":{"aider-polyglot":48.4,"gpqa-diamond":56.5,"hellaswag":88.9,"math-level-5":64.9,"metr-time-horizon":18.47,"mmlu":87.2,"winogrande":85.2},"weights":"open"},{"access":"API access","benchmarks":8,"country":"United States of America","eci":141.9,"eci_high":143.2,"eci_low":139.7,"eci_rank":106,"key":"o1","name":"o1","org":"OpenAI","org_first":"OpenAI","released":"2024-12-17","scores":{"aider-polyglot":61.7,"arc-agi":30.7,"frontiermath":14.7,"gpqa-diamond":76.8,"hle":8,"math-level-5":94.7,"metr-time-horizon":38.83,"simpleqa-verified":41.1},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":3,"country":"United States of America","eci":130.4,"eci_high":132,"eci_low":124.7,"eci_rank":157,"key":"phi-4","name":"Phi-4","org":"Microsoft Research","org_first":"Microsoft Research","released":"2024-12-12","scores":{"gpqa-diamond":56.1,"math-level-5":64.9,"mmlu":84.8},"weights":"open"},{"access":"API access","benchmarks":2,"country":"United States of America","eci":130.5,"eci_high":132.2,"eci_low":125.8,"eci_rank":156,"key":"grok-2-dec-2024","name":"Grok-2 (Dec 2024)","org":"xAI","org_first":"xAI","released":"2024-12-12","scores":{"gpqa-diamond":53.8,"math-level-5":63.5},"weights":"closed"},{"access":"Open weights (restricted use)","benchmarks":3,"country":"United States of America","eci":127.3,"eci_high":129.7,"eci_low":121.8,"eci_rank":172,"key":"llama-3-3-70b","name":"Llama 3.3 70B","org":"Meta AI","org_first":"Meta AI","released":"2024-12-06","scores":{"gpqa-diamond":47.4,"math-level-5":41.6,"mmlu":86.3},"weights":"open"},{"access":"API access","benchmarks":2,"country":"United States of America","eci":134.7,"eci_high":136.6,"eci_low":124.9,"eci_rank":144,"key":"gemini-2-0-flash-dec-2024","name":"Gemini 2.0 Flash (Dec 2024)","org":"Google DeepMind, Google","org_first":"Google DeepMind","released":"2024-12-06","scores":{"aider-polyglot":38.2,"mmlu":79.7},"weights":"closed"},{"access":"API access","benchmarks":2,"country":"United States of America","eci":123.8,"eci_high":126.8,"eci_low":110.5,"eci_rank":185,"key":"amazon-nova-pro","name":"Amazon Nova Pro","org":"Amazon","org_first":"Amazon","released":"2024-12-03","scores":{"hle":4.4,"mmlu":82},"weights":"closed"},{"access":"API access","benchmarks":1,"country":"United States of America","key":"amazon-nova-micro","name":"Amazon Nova Micro","org":"Amazon","org_first":"Amazon","released":"2024-12-03","scores":{"mmlu":70.8},"weights":"closed"},{"access":"API access","benchmarks":2,"country":"United States of America","key":"amazon-nova-lite","name":"Amazon Nova Lite","org":"Amazon","org_first":"Amazon","released":"2024-12-03","scores":{"hle":3.6,"mmlu":77},"weights":"closed"},{"access":"API access","benchmarks":1,"country":"China","key":"yi-lightning","name":"Yi-Lightning","org":"01.AI","org_first":"01.AI","released":"2024-12-02","scores":{"aider-polyglot":12.9},"weights":"closed"},{"access":"Unknown","benchmarks":4,"country":"United States of America","eci":100.9,"eci_high":106.3,"eci_low":88.5,"eci_rank":239,"key":"intellect-1","name":"INTELLECT-1","org":"Prime Intellect, Hugging Face, Arcee AI","org_first":"Prime Intellect","released":"2024-11-29","scores":{"gsm8k":38.6,"hellaswag":71.4,"mmlu":49.9,"winogrande":65.8},"weights":"unknown"},{"access":"Open weights (restricted use)","benchmarks":2,"country":"United States of America","key":"tulu-3-t-lu-3-70b","name":"Tulu 3 (Tülu 3) 70B","org":"Allen Institute for AI, University of Washington","org_first":"Allen Institute for AI","released":"2024-11-21","scores":{"gpqa-diamond":46.3,"math-level-5":42.7},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":2,"country":"China","key":"qwen2-5-coder-32b-instruct","name":"Qwen2.5-Coder-32B-Instruct","org":"Alibaba","org_first":"Alibaba","released":"2024-11-21","scores":{"aider-polyglot":16.4,"gsm8k":93},"weights":"open"},{"access":"API access","benchmarks":8,"country":"United States of America","eci":128.8,"eci_high":131.2,"eci_low":123.6,"eci_rank":167,"key":"gpt-4o-nov-2024","name":"GPT-4o (Nov 2024)","org":"OpenAI","org_first":"OpenAI","released":"2024-11-20","scores":{"aider-polyglot":18.2,"arc-agi":4.5,"arc-agi-2":0,"gpqa-diamond":47.9,"hle":2.7,"math-level-5":49.8,"mmlu":88.1,"swe-bench-verified":31},"weights":"closed"},{"access":"Open weights (non-commercial)","benchmarks":2,"country":"France","eci":128.5,"eci_high":130.6,"eci_low":122.7,"eci_rank":169,"key":"mistral-large-2-nov-2024","name":"Mistral Large 2 (Nov 2024)","org":"Mistral AI","org_first":"Mistral AI","released":"2024-11-15","scores":{"gpqa-diamond":51.3,"math-level-5":50.3},"weights":"open"},{"access":"API access","benchmarks":2,"country":"China","key":"qwen-turbo","name":"Qwen-Turbo","org":"Alibaba","org_first":"Alibaba","released":"2024-11-01","scores":{"gpqa-diamond":41.8,"math-level-5":56.2},"weights":"closed"},{"access":"API access","benchmarks":6,"country":"United States of America","eci":133.6,"eci_high":137.6,"eci_low":128.8,"eci_rank":148,"key":"claude-3-5-sonnet-october-2024","name":"Claude 3.5 Sonnet (October 2024)","org":"Anthropic","org_first":"Anthropic","released":"2024-10-22","scores":{"aider-polyglot":51.6,"gpqa-diamond":55.3,"hle":4.1,"math-level-5":56.9,"metr-time-horizon":20.52,"mmlu":87.3},"weights":"closed"},{"access":"API access","benchmarks":4,"country":"United States of America","eci":127.2,"eci_high":129.5,"eci_low":120.7,"eci_rank":174,"key":"claude-3-5-haiku","name":"Claude 3.5 Haiku","org":"Anthropic","org_first":"Anthropic","released":"2024-10-22","scores":{"aider-polyglot":28,"gpqa-diamond":38.1,"math-level-5":46.4,"mmlu":74.3},"weights":"closed"},{"access":"Open weights (non-commercial)","benchmarks":2,"country":"France","key":"ministral-8b","name":"Ministral 8B","org":"Mistral AI","org_first":"Mistral AI","released":"2024-10-16","scores":{"gpqa-diamond":27.1,"math-level-5":14.9},"weights":"open"},{"access":"API access","benchmarks":2,"country":"France","eci":118.1,"eci_high":121.7,"eci_low":106.2,"eci_rank":207,"key":"ministral-3b","name":"Ministral 3B","org":"Mistral AI","org_first":"Mistral AI","released":"2024-10-16","scores":{"gpqa-diamond":25.3,"math-level-5":14.4},"weights":"closed"},{"access":"API access","benchmarks":1,"country":"United States of America","key":"gemini-1-5-flash-8b","name":"Gemini 1.5 Flash 8B","org":"Google DeepMind","org_first":"Google DeepMind","released":"2024-10-03","scores":{"gpqa-diamond":33},"weights":"closed"},{"access":"Open weights (restricted use)","benchmarks":3,"country":"United States of America","eci":125.5,"eci_high":127.3,"eci_low":119.4,"eci_rank":182,"key":"llama-3-2-90b","name":"Llama 3.2 90B","org":"Meta AI","org_first":"Meta AI","released":"2024-09-24","scores":{"gpqa-diamond":41,"math-level-5":39.4,"mmlu":80.3},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":1,"country":"United States of America","eci":102,"eci_high":110.1,"eci_low":90,"eci_rank":238,"key":"llama-3-2-1b","name":"Llama 3.2 1B","org":"Meta AI","org_first":"Meta AI","released":"2024-09-24","scores":{"gpqa-diamond":23.9},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":1,"country":"United States of America","key":"llama-3-2-11b","name":"Llama 3.2 11B","org":"Meta AI","org_first":"Meta AI","released":"2024-09-24","scores":{"mmlu":56.5},"weights":"open"},{"access":"API access","benchmarks":5,"country":"United States of America","eci":131.7,"eci_high":133.1,"eci_low":127.2,"eci_rank":153,"key":"gemini-1-5-pro-sept-2024","name":"Gemini 1.5 Pro (Sept 2024)","org":"Google DeepMind","org_first":"Google DeepMind","released":"2024-09-24","scores":{"arc-agi-2":0.8,"gpqa-diamond":57.2,"hle":4.6,"math-level-5":70.4,"mmlu":86.9},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":2,"country":"China","eci":118.5,"eci_high":121.2,"eci_low":110.9,"eci_rank":205,"key":"qwen2-5-7b","name":"Qwen2.5-7B","org":"Alibaba","org_first":"Alibaba","released":"2024-09-19","scores":{"gpqa-diamond":35.5,"mmlu":72.9},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":6,"country":"China","eci":129,"eci_high":130.6,"eci_low":123.2,"eci_rank":164,"key":"qwen2-5-72b","name":"Qwen2.5-72B","org":"Alibaba","org_first":"Alibaba","released":"2024-09-19","scores":{"gpqa-diamond":49.1,"hellaswag":84.8,"math-level-5":63.2,"metr-time-horizon":5.17,"mmlu":85.3,"winogrande":82.3},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":1,"country":"China","key":"qwen2-5-14b","name":"Qwen2.5-14B","org":"Alibaba","org_first":"Alibaba","released":"2024-09-19","scores":{"mmlu":79.9},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":4,"country":"China","eci":119.5,"eci_high":124.9,"eci_low":110,"eci_rank":199,"key":"qwen2-5-coder-32b","name":"Qwen2.5-Coder-32B","org":"Alibaba","org_first":"Alibaba","released":"2024-09-18","scores":{"gsm8k":91.1,"hellaswag":83,"mmlu":79.1,"winogrande":80.8},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":4,"country":"China","eci":113.1,"eci_high":119.9,"eci_low":102.6,"eci_rank":218,"key":"qwen2-5-coder-7b","name":"Qwen2.5-Coder (7B)","org":"Alibaba","org_first":"Alibaba","released":"2024-09-18","scores":{"gsm8k":86.7,"hellaswag":76.8,"mmlu":68,"winogrande":72.9},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":4,"country":"China","eci":103,"eci_high":111,"eci_low":85.3,"eci_rank":237,"key":"qwen2-5-coder-1-5b","name":"Qwen2.5-Coder (1.5B)","org":"Alibaba","org_first":"Alibaba","released":"2024-09-18","scores":{"gsm8k":65.8,"hellaswag":61.8,"mmlu":53.6,"winogrande":60.7},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":2,"country":"China","eci":128.5,"eci_high":130,"eci_low":122.2,"eci_rank":168,"key":"qwen2-5-32b","name":"Qwen2.5-32B","org":"Alibaba","org_first":"Alibaba","released":"2024-09-17","scores":{"gpqa-diamond":46.1,"math-level-5":56.1},"weights":"open"},{"access":"API access","benchmarks":4,"country":"United States of America","eci":134.8,"eci_high":138.7,"eci_low":130.8,"eci_rank":143,"key":"o1-preview","name":"o1-preview","org":"OpenAI","org_first":"OpenAI","released":"2024-09-12","scores":{"arc-agi":18,"gpqa-diamond":50.3,"math-level-5":81.6,"metr-time-horizon":20.33},"weights":"closed"},{"access":"API access","benchmarks":5,"country":"United States of America","eci":135.8,"eci_high":137.3,"eci_low":131.7,"eci_rank":138,"key":"o1-mini","name":"o1-mini","org":"OpenAI","org_first":"OpenAI","released":"2024-09-12","scores":{"aider-polyglot":32.9,"arc-agi":14,"arc-agi-2":0.8,"gpqa-diamond":62.4,"math-level-5":89.2},"weights":"closed"},{"access":"Open weights (restricted use)","benchmarks":1,"country":"China","key":"deepseek-v2-5-sep-2024","name":"DeepSeek-V2.5 (Sep 2024)","org":"DeepSeek","org_first":"DeepSeek","released":"2024-09-06","scores":{"aider-polyglot":17.8},"weights":"open"},{"access":"Open weights (non-commercial)","benchmarks":1,"country":"Canada","eci":119.3,"eci_high":125.3,"eci_low":109.9,"eci_rank":201,"key":"command-r","name":"Command R+","org":"Cohere, Cohere Labs (formerly Cohere for AI)","org_first":"Cohere","released":"2024-08-30","scores":{"mmlu":69.4},"weights":"open"},{"access":"API access","benchmarks":3,"country":"United States of America","eci":129.4,"eci_high":131.2,"eci_low":123.2,"eci_rank":162,"key":"gemini-1-5-flash-sep-2024","name":"Gemini 1.5 Flash (Sep 2024)","org":"Google DeepMind","org_first":"Google DeepMind","released":"2024-08-27","scores":{"gpqa-diamond":47.3,"math-level-5":61.9,"mmlu":73.9},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":1,"country":"United States of America","key":"phi-3-5-moe","name":"Phi-3.5-MoE","org":"Microsoft","org_first":"Microsoft","released":"2024-08-17","scores":{"gsm8k":88.7},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":1,"country":"United States of America","key":"phi-3-5-mini","name":"Phi-3.5-mini","org":"Microsoft","org_first":"Microsoft","released":"2024-08-16","scores":{"gsm8k":86.2},"weights":"open"},{"access":"API access","benchmarks":6,"country":"United States of America","eci":128.8,"eci_high":130.7,"eci_low":123.6,"eci_rank":166,"key":"gpt-4o-aug-2024","name":"GPT-4o (Aug 2024)","org":"OpenAI","org_first":"OpenAI","released":"2024-08-06","scores":{"aider-polyglot":23.1,"frontiermath":0.4,"gpqa-diamond":49.2,"math-level-5":53.3,"mmlu":84.3,"simpleqa-verified":26},"weights":"closed"},{"access":"Open weights (non-commercial)","benchmarks":3,"country":"France","eci":127.5,"eci_high":129.6,"eci_low":122,"eci_rank":170,"key":"mistral-large-2-jul-2024","name":"Mistral Large 2 (Jul 2024)","org":"Mistral AI","org_first":"Mistral AI","released":"2024-07-24","scores":{"gpqa-diamond":49,"math-level-5":44.8,"mmlu":80},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":4,"country":"United States of America","eci":116.6,"eci_high":121.1,"eci_low":106.2,"eci_rank":212,"key":"llama-3-1-8b","name":"Llama 3.1-8B","org":"Meta AI","org_first":"Meta AI","released":"2024-07-23","scores":{"gpqa-diamond":27,"gsm8k":82.4,"math-level-5":22.9,"mmlu":56.1},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":3,"country":"United States of America","eci":125.9,"eci_high":128.5,"eci_low":120,"eci_rank":181,"key":"llama-3-1-70b","name":"Llama 3.1-70B","org":"Meta AI","org_first":"Meta AI","released":"2024-07-23","scores":{"gpqa-diamond":44.2,"math-level-5":36.7,"mmlu":80.1},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":5,"country":"United States of America","eci":128.8,"eci_high":130.8,"eci_low":123.8,"eci_rank":165,"key":"llama-3-1-405b","name":"Llama 3.1-405B","org":"Meta AI","org_first":"Meta AI","released":"2024-07-23","scores":{"gpqa-diamond":50.9,"hellaswag":89.2,"math-level-5":49.8,"mmlu":84.5,"winogrande":89.2},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":3,"country":"France","eci":118.7,"eci_high":121.2,"eci_low":111,"eci_rank":202,"key":"mistral-nemo","name":"Mistral NeMo","org":"Mistral AI","org_first":"Mistral AI","released":"2024-07-18","scores":{"gpqa-diamond":29.9,"gsm8k":84.2,"math-level-5":10.8},"weights":"open"},{"access":"API access","benchmarks":8,"country":"United States of America","eci":126.6,"eci_high":128.7,"eci_low":119.1,"eci_rank":178,"key":"gpt-4o-mini","name":"GPT-4o mini","org":"OpenAI","org_first":"OpenAI","released":"2024-07-18","scores":{"aider-polyglot":3.6,"arc-agi-2":0,"frontiermath":0.7,"gpqa-diamond":37.7,"gsm8k":91.3,"math-level-5":52.6,"mmlu":81.8,"simpleqa-verified":8.3},"weights":"closed"},{"access":"Open weights (restricted use)","benchmarks":4,"country":"United States of America","eci":119.8,"eci_high":122.6,"eci_low":111.6,"eci_rank":198,"key":"gemma-2-9b","name":"Gemma 2 9B","org":"Google DeepMind","org_first":"Google DeepMind","released":"2024-06-24","scores":{"gpqa-diamond":27.5,"gsm8k":84.9,"math-level-5":21,"mmlu":72.1},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":3,"country":"United States of America","eci":122.1,"eci_high":124.1,"eci_low":114.4,"eci_rank":190,"key":"gemma-2-27b","name":"Gemma 2 27B","org":"Google DeepMind","org_first":"Google DeepMind","released":"2024-06-24","scores":{"gpqa-diamond":36.5,"math-level-5":27.9,"mmlu":75.7},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":2,"country":"United States of America","key":"hermes-2-theta-llama-3-70b","name":"Hermes 2 Theta Llama-3 70B","org":"Nous Research, Arcee AI","org_first":"Nous Research","released":"2024-06-20","scores":{"gpqa-diamond":37.5,"math-level-5":22.7},"weights":"open"},{"access":"API access","benchmarks":4,"country":"United States of America","eci":130,"eci_rank":158,"key":"claude-3-5-sonnet","name":"Claude 3.5 Sonnet","org":"Anthropic","org_first":"Anthropic","released":"2024-06-20","scores":{"gpqa-diamond":54,"math-level-5":51.7,"metr-time-horizon":11.4,"mmlu":86.5},"weights":"closed"},{"access":"Open weights (restricted use)","benchmarks":2,"country":"China","key":"deepseek-coder-v2-236b","name":"DeepSeek-Coder-V2 236B","org":"DeepSeek","org_first":"DeepSeek","released":"2024-06-17","scores":{"gsm8k":94.5,"winogrande":83.7},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":4,"country":"China","eci":125.3,"eci_high":126.9,"eci_low":119,"eci_rank":183,"key":"qwen2-72b","name":"Qwen2-72B","org":"Alibaba","org_first":"Alibaba","released":"2024-06-07","scores":{"gpqa-diamond":40.8,"math-level-5":39.1,"metr-time-horizon":2.24,"mmlu":82.4},"weights":"open"},{"access":"API access","benchmarks":3,"country":"United States of America","eci":126.9,"eci_high":129.2,"eci_low":121.1,"eci_rank":177,"key":"gemini-1-5-pro-may-2024","name":"Gemini 1.5 Pro (May 2024)","org":"Google DeepMind","org_first":"Google DeepMind","released":"2024-05-14","scores":{"gpqa-diamond":45.9,"math-level-5":40.7,"mmlu":85.9},"weights":"closed"},{"access":"API access","benchmarks":4,"country":"United States of America","eci":122.6,"eci_high":124.6,"eci_low":115.7,"eci_rank":189,"key":"gemini-1-5-flash-may-2024","name":"Gemini 1.5 Flash (May 2024)","org":"Google DeepMind","org_first":"Google DeepMind","released":"2024-05-14","scores":{"gpqa-diamond":40.4,"gsm8k":82.4,"math-level-5":25.1,"mmlu":77.9},"weights":"closed"},{"access":"API access","benchmarks":1,"country":"China","key":"yi-large","name":"Yi-Large","org":"01.AI","org_first":"01.AI","released":"2024-05-13","scores":{"mmlu":79.3},"weights":"closed"},{"access":"Open weights (restricted use)","benchmarks":2,"country":"China","key":"yi-1-5-34b","name":"Yi-1.5-34B","org":"01.AI","org_first":"01.AI","released":"2024-05-13","scores":{"gpqa-diamond":32,"math-level-5":25.5},"weights":"open"},{"access":"API access","benchmarks":4,"country":"United States of America","eci":129,"eci_high":131.7,"eci_low":123.8,"eci_rank":163,"key":"gpt-4o-may-2024","name":"GPT-4o (May 2024)","org":"OpenAI","org_first":"OpenAI","released":"2024-05-13","scores":{"gpqa-diamond":48.9,"math-level-5":51,"metr-time-horizon":9.17,"mmlu":84.2},"weights":"closed"},{"access":"Open weights (restricted use)","benchmarks":4,"country":"United Arab Emirates","eci":109.6,"eci_high":115.6,"eci_low":99.8,"eci_rank":225,"key":"falcon-2-11b","name":"Falcon 2 11B","org":"Technology Innovation Institute","org_first":"Technology Innovation Institute","released":"2024-05-09","scores":{"gsm8k":53.8,"hellaswag":82.9,"mmlu":58.4,"winogrande":78.3},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":3,"country":"China","eci":124.8,"eci_high":127.8,"eci_low":118.5,"eci_rank":184,"key":"deepseek-v2-moe-236b-may-2024","name":"DeepSeek-V2 (MoE-236B, May 2024)","org":"DeepSeek","org_first":"DeepSeek","released":"2024-05-07","scores":{"hellaswag":87.1,"mmlu":78.4,"winogrande":86.3},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":3,"country":"United States of America","eci":121.8,"eci_high":124.4,"eci_low":114.8,"eci_rank":193,"key":"phi-3-small-7-4b","name":"phi-3-small 7.4B","org":"Microsoft","org_first":"Microsoft","released":"2024-04-23","scores":{"hellaswag":77,"mmlu":75.7,"winogrande":81.5},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":3,"country":"United States of America","eci":117.4,"eci_high":120.7,"eci_low":107.9,"eci_rank":209,"key":"phi-3-mini-3-8b","name":"phi-3-mini 3.8B","org":"Microsoft","org_first":"Microsoft","released":"2024-04-23","scores":{"hellaswag":76.7,"mmlu":68.8,"winogrande":70.8},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":5,"country":"United States of America","eci":121.2,"eci_high":124.1,"eci_low":113.8,"eci_rank":194,"key":"phi-3-medium-14b","name":"phi-3-medium 14B","org":"Microsoft","org_first":"Microsoft","released":"2024-04-23","scores":{"gpqa-diamond":27.6,"hellaswag":82.4,"math-level-5":17.6,"mmlu":78,"winogrande":81.5},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":4,"country":"United States of America","eci":116.5,"eci_high":119.3,"eci_low":106.8,"eci_rank":213,"key":"llama-3-8b","name":"Llama 3-8B","org":"Meta AI","org_first":"Meta AI","released":"2024-04-18","scores":{"gpqa-diamond":26.1,"math-level-5":6.1,"mmlu":68.8,"winogrande":75.7},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":4,"country":"United States of America","eci":122.9,"eci_high":125.8,"eci_low":116.6,"eci_rank":188,"key":"llama-3-70b","name":"Llama 3-70B","org":"Meta AI","org_first":"Meta AI","released":"2024-04-18","scores":{"gpqa-diamond":40.6,"math-level-5":22.6,"mmlu":79.3,"winogrande":83.5},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":3,"country":"France","eci":122,"eci_high":124.5,"eci_low":114.6,"eci_rank":192,"key":"mixtral-8x22b","name":"Mixtral 8x22B","org":"Mistral AI","org_first":"Mistral AI","released":"2024-04-17","scores":{"gpqa-diamond":34.1,"math-level-5":24.2,"mmlu":77.8},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":2,"country":"United States of America","key":"wizardlm-2-8x22b","name":"WizardLM-2 8x22B","org":"Microsoft","org_first":"Microsoft","released":"2024-04-15","scores":{"gpqa-diamond":43.4,"math-level-5":25.7},"weights":"open"},{"access":"API access","benchmarks":5,"country":"United States of America","eci":127.2,"eci_high":128.8,"eci_low":121.9,"eci_rank":173,"key":"gpt-4-turbo-apr-2024","name":"GPT-4 Turbo (Apr 2024)","org":"OpenAI","org_first":"OpenAI","released":"2024-04-09","scores":{"frontiermath":0.7,"gpqa-diamond":46.6,"math-level-5":46.7,"metr-time-horizon":6.57,"mmlu":81.3},"weights":"closed"},{"access":"Open weights (restricted use)","benchmarks":2,"country":"United States of America","key":"dbrx","name":"DBRX","org":"Databricks","org_first":"Databricks","released":"2024-03-27","scores":{"gpqa-diamond":32.9,"math-level-5":11.7},"weights":"open"},{"access":"API access","benchmarks":4,"country":"United States of America","eci":118.3,"eci_high":121.5,"eci_low":110.5,"eci_rank":206,"key":"claude-3-haiku","name":"Claude 3 Haiku","org":"Anthropic","org_first":"Anthropic","released":"2024-03-07","scores":{"gpqa-diamond":36.3,"math-level-5":14.9,"mmlu":73.8,"winogrande":74.2},"weights":"closed"},{"access":"API access","benchmarks":4,"country":"United States of America","eci":120.7,"eci_high":123.9,"eci_low":113.2,"eci_rank":195,"key":"claude-3-sonnet","name":"Claude 3 Sonnet","org":"Anthropic","org_first":"Anthropic","released":"2024-02-29","scores":{"gpqa-diamond":40.6,"math-level-5":18.2,"mmlu":75.9,"winogrande":75.1},"weights":"closed"},{"access":"API access","benchmarks":6,"country":"United States of America","eci":126.9,"eci_high":130,"eci_low":121.5,"eci_rank":176,"key":"claude-3-opus","name":"Claude 3 Opus","org":"Anthropic","org_first":"Anthropic","released":"2024-02-29","scores":{"gpqa-diamond":47.2,"math-level-5":37.5,"metr-time-horizon":3.95,"mmlu":84.6,"simpleqa-verified":12.6,"winogrande":88.5},"weights":"closed"},{"access":"Unreleased","benchmarks":4,"country":"United States of America","eci":107.7,"eci_high":113.2,"eci_low":97.2,"eci_rank":228,"key":"nemotron-4-15b","name":"Nemotron-4 15B","org":"Nvidia","org_first":"Nvidia","released":"2024-02-26","scores":{"gsm8k":46,"hellaswag":82.4,"mmlu":58.7,"winogrande":78},"weights":"closed"},{"access":"API access","benchmarks":3,"country":"France","eci":122,"eci_high":125.1,"eci_low":114.4,"eci_rank":191,"key":"mistral-large","name":"Mistral Large","org":"Mistral AI","org_first":"Mistral AI","released":"2024-02-26","scores":{"gpqa-diamond":38.8,"math-level-5":24.5,"mmlu":68.8},"weights":"closed"},{"access":"Open weights (restricted use)","benchmarks":3,"country":"United States of America","eci":88.8,"eci_high":97.5,"eci_low":70.7,"eci_rank":255,"key":"starcoder-2-3b","name":"StarCoder 2 3B","org":"Hugging Face, ServiceNow, Nvidia, BigCode","org_first":"Hugging Face","released":"2024-02-22","scores":{"gsm8k":21.6,"mmlu":36.6,"winogrande":57.1},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":4,"country":"United States of America","eci":112,"eci_high":116.5,"eci_low":102.4,"eci_rank":222,"key":"gemma-7b","name":"Gemma 7B","org":"Google DeepMind","org_first":"Google DeepMind","released":"2024-02-21","scores":{"gsm8k":46.4,"hellaswag":82.2,"mmlu":66.1,"winogrande":79},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":4,"country":"United States of America","eci":94.2,"eci_high":101.3,"eci_low":80.8,"eci_rank":248,"key":"gemma-2b","name":"Gemma 2B","org":"Google DeepMind","org_first":"Google DeepMind","released":"2024-02-21","scores":{"gsm8k":17.7,"hellaswag":71.4,"mmlu":42.3,"winogrande":65.4},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":3,"country":"United States of America","eci":93.6,"eci_high":101.7,"eci_low":76.8,"eci_rank":249,"key":"starcoder-2-7b","name":"StarCoder 2 7B","org":"Hugging Face, ServiceNow, Nvidia, BigCode","org_first":"Hugging Face","released":"2024-02-20","scores":{"gsm8k":32.7,"mmlu":38.8,"winogrande":57.1},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":3,"country":"United States of America","eci":105,"eci_high":111.3,"eci_low":91,"eci_rank":233,"key":"starcoder-2-15b","name":"StarCoder 2 15B","org":"Hugging Face, ServiceNow, Nvidia, BigCode","org_first":"Hugging Face","released":"2024-02-20","scores":{"gsm8k":57.7,"mmlu":64.1,"winogrande":64.3},"weights":"open"},{"access":"API access","benchmarks":1,"country":"United States of America","key":"gemini-1-5-pro-feb-2024","name":"Gemini 1.5 Pro (Feb 2024)","org":"Google DeepMind","org_first":"Google DeepMind","released":"2024-02-15","scores":{"mmlu":82.7},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":1,"country":"China","key":"qwen1-5-7b","name":"Qwen1.5-7B","org":"Alibaba","org_first":"Alibaba","released":"2024-02-04","scores":{"mmlu":62.6},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":1,"country":"China","key":"qwen1-5-72b","name":"Qwen1.5-72B","org":"Alibaba","org_first":"Alibaba","released":"2024-02-04","scores":{"gpqa-diamond":28.8},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":2,"country":"China","key":"qwen1-5-32b","name":"Qwen1.5-32B","org":"Alibaba","org_first":"Alibaba","released":"2024-02-04","scores":{"gpqa-diamond":30.7,"mmlu":74.4},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":1,"country":"China","key":"qwen1-5-14b","name":"Qwen1.5-14B","org":"Alibaba","org_first":"Alibaba","released":"2024-02-04","scores":{"mmlu":68.6},"weights":"open"},{"access":"API access","benchmarks":4,"country":"United States of America","eci":115.7,"eci_high":118.9,"eci_low":107.4,"eci_rank":215,"key":"gpt-3-5-turbo-jan-2024","name":"GPT-3.5 Turbo (Jan 2024)","org":"OpenAI","org_first":"OpenAI","released":"2024-01-25","scores":{"frontiermath":0,"gpqa-diamond":27.2,"math-level-5":11.6,"mmlu":67.3},"weights":"closed"},{"access":"API access","benchmarks":3,"country":"United States of America","eci":117,"eci_high":119.5,"eci_low":108.8,"eci_rank":211,"key":"gemini-1-0-pro","name":"Gemini 1.0 Pro","org":"Google DeepMind","org_first":"Google DeepMind","released":"2023-12-13","scores":{"gpqa-diamond":34,"math-level-5":11.2,"mmlu":70},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":3,"country":"United States of America","eci":107.9,"eci_high":113.8,"eci_low":88.6,"eci_rank":227,"key":"phi-2","name":"Phi-2","org":"Microsoft","org_first":"Microsoft","released":"2023-12-12","scores":{"hellaswag":53.6,"mmlu":58.4,"winogrande":54.7},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":6,"country":"France","eci":118.5,"eci_high":121.1,"eci_low":110.4,"eci_rank":204,"key":"mixtral-8x7b","name":"Mixtral 8x7B","org":"Mistral AI","org_first":"Mistral AI","released":"2023-12-11","scores":{"gpqa-diamond":30.6,"gsm8k":74.4,"hellaswag":86.7,"math-level-5":10,"mmlu":70.6,"winogrande":77.2},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":2,"country":"France","key":"mistral-7b-v0-2","name":"Mistral 7B v0.2","org":"Mistral AI","org_first":"Mistral AI","released":"2023-12-11","scores":{"gsm8k":35.4,"mmlu":62.5},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":2,"country":"China","eci":110.5,"eci_high":116.6,"eci_low":95.6,"eci_rank":223,"key":"deepseek-llm-67b","name":"DeepSeek LLM 67B","org":"DeepSeek","org_first":"DeepSeek","released":"2023-11-29","scores":{"gpqa-diamond":24.6,"math-level-5":6.4},"weights":"open"},{"access":"API access","benchmarks":2,"country":"United States of America","eci":119.3,"eci_high":122,"eci_low":111.3,"eci_rank":200,"key":"claude-2-1","name":"Claude 2.1","org":"Anthropic","org_first":"Anthropic","released":"2023-11-21","scores":{"gpqa-diamond":33,"mmlu":73.5},"weights":"closed"},{"access":"API access","benchmarks":4,"country":"United States of America","eci":126.5,"eci_high":129,"eci_low":120.9,"eci_rank":179,"key":"gpt-4-turbo-nov-2023","name":"GPT-4 Turbo (Nov 2023)","org":"OpenAI","org_first":"OpenAI","released":"2023-11-06","scores":{"gpqa-diamond":42.4,"math-level-5":40,"metr-time-horizon":5.37,"mmlu":79.6},"weights":"closed"},{"access":"API access","benchmarks":4,"country":"United States of America","eci":118.5,"eci_high":121.7,"eci_low":110,"eci_rank":203,"key":"gpt-3-5-turbo-nov-2023","name":"GPT-3.5 Turbo (Nov 2023)","org":"OpenAI","org_first":"OpenAI","released":"2023-11-06","scores":{"gpqa-diamond":28,"math-level-5":15.9,"mmlu":71.4,"winogrande":68.8},"weights":"closed"},{"access":"Open weights (restricted use)","benchmarks":4,"country":"China","eci":117.4,"eci_high":120.6,"eci_low":108,"eci_rank":208,"key":"yi-34b","name":"Yi-34B","org":"01.AI","org_first":"01.AI","released":"2023-11-02","scores":{"gpqa-diamond":14.7,"gsm8k":76,"math-level-5":5.1,"mmlu":76.3},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":4,"country":"China","eci":104.8,"eci_high":110.3,"eci_low":94.4,"eci_rank":234,"key":"yi-6b","name":"Yi 6B","org":"01.AI","org_first":"01.AI","released":"2023-11-02","scores":{"gsm8k":44.9,"hellaswag":74.4,"mmlu":64,"winogrande":71.3},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":3,"country":"China","eci":89.6,"eci_high":97.8,"eci_low":73,"eci_rank":254,"key":"deepseek-coder-6-7b","name":"DeepSeek Coder 6.7B","org":"DeepSeek, Peking University","org_first":"DeepSeek","released":"2023-11-02","scores":{"gsm8k":21.3,"mmlu":36.4,"winogrande":57.6},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":3,"country":"China","eci":96.3,"eci_high":103.5,"eci_low":82.8,"eci_rank":245,"key":"deepseek-coder-33b","name":"DeepSeek Coder 33B","org":"DeepSeek, Peking University","org_first":"DeepSeek","released":"2023-11-02","scores":{"gsm8k":35.4,"mmlu":39.4,"winogrande":62},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":3,"country":"China","eci":63.6,"eci_high":76.8,"eci_low":42.3,"eci_rank":257,"key":"deepseek-coder-1-3b","name":"DeepSeek Coder 1.3B","org":"DeepSeek, Peking University","org_first":"DeepSeek","released":"2023-11-02","scores":{"gsm8k":4.4,"mmlu":25.8,"winogrande":53.3},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":2,"country":"China","eci":106.8,"eci_high":111.9,"eci_low":93.6,"eci_rank":230,"key":"qwen-7b","name":"Qwen-7B","org":"Alibaba","org_first":"Alibaba","released":"2023-09-28","scores":{"gsm8k":51.7,"mmlu":45},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":3,"country":"France","eci":109,"eci_high":112.5,"eci_low":96.8,"eci_rank":226,"key":"mistral-7b-v0-3","name":"Mistral 7B v0.3","org":"Mistral AI","org_first":"Mistral AI","released":"2023-09-27","scores":{"gpqa-diamond":15.2,"math-level-5":3.7,"mmlu":59.9},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":4,"country":"France","eci":112.2,"eci_high":116.3,"eci_low":102.4,"eci_rank":220,"key":"mistral-7b-v0-1","name":"Mistral 7B v0.1","org":"Mistral AI","org_first":"Mistral AI","released":"2023-09-27","scores":{"gsm8k":54.4,"hellaswag":81,"mmlu":62.5,"winogrande":75.3},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":2,"country":"China","eci":113,"eci_high":117.4,"eci_low":103.3,"eci_rank":219,"key":"qwen-14b","name":"Qwen-14B","org":"Alibaba","org_first":"Alibaba","released":"2023-09-24","scores":{"gsm8k":61.3,"mmlu":66.3},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":3,"country":"China","eci":96.3,"eci_high":104.3,"eci_low":83.8,"eci_rank":244,"key":"baichuan-2-7b","name":"Baichuan 2-7B","org":"Baichuan","org_first":"Baichuan","released":"2023-09-20","scores":{"gsm8k":24.6,"hellaswag":68,"mmlu":54.2},"weights":"open"},{"access":"API access","benchmarks":1,"country":"United States of America","key":"gpt-3-5-turbo-instruct","name":"GPT-3.5 Turbo Instruct","org":"OpenAI","org_first":"OpenAI","released":"2023-09-18","scores":{"metr-time-horizon":0.6},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":3,"country":"United States of America","eci":91.5,"eci_high":100.8,"eci_low":71.6,"eci_rank":251,"key":"phi-1-5","name":"Phi-1.5","org":"Microsoft","org_first":"Microsoft","released":"2023-09-11","scores":{"hellaswag":47.6,"mmlu":37.6,"winogrande":73.4},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":4,"country":"United Arab Emirates","eci":112.1,"eci_high":118.9,"eci_low":103.1,"eci_rank":221,"key":"falcon-180b","name":"Falcon-180B","org":"Technology Innovation Institute","org_first":"Technology Innovation Institute","released":"2023-09-06","scores":{"gsm8k":54.4,"hellaswag":89,"mmlu":70.6,"winogrande":87.1},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":3,"country":"China","eci":103.2,"eci_high":110.3,"eci_low":91.1,"eci_rank":236,"key":"baichuan2-13b","name":"Baichuan2-13B","org":"Baichuan","org_first":"Baichuan","released":"2023-09-06","scores":{"gsm8k":52.8,"hellaswag":70.8,"mmlu":59.2},"weights":"open"},{"access":"API access","benchmarks":2,"country":"United States of America","eci":120.3,"eci_high":122.8,"eci_low":112.5,"eci_rank":196,"key":"claude-instant","name":"Claude Instant","org":"Anthropic","org_first":"Anthropic","released":"2023-08-09","scores":{"gsm8k":86.7,"mmlu":73.4},"weights":"closed"},{"access":"Open weights (non-commercial)","benchmarks":3,"country":"United Kingdom","eci":117.1,"eci_high":120.1,"eci_low":108.4,"eci_rank":210,"key":"stable-beluga-2","name":"Stable Beluga 2","org":"Stability AI","org_first":"Stability AI","released":"2023-07-20","scores":{"gsm8k":69.6,"hellaswag":84.1,"mmlu":68.6},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":4,"country":"United States of America","eci":99.1,"eci_high":105.6,"eci_low":86.8,"eci_rank":242,"key":"llama-2-7b","name":"Llama 2-7B","org":"Meta AI","org_first":"Meta AI","released":"2023-07-18","scores":{"gsm8k":16.7,"hellaswag":77.2,"mmlu":45.8,"winogrande":69.2},"weights":"open"},{"access":"Open weights (restricted use)","benchmarks":6,"country":"United States of America","eci":113.8,"eci_high":117.4,"eci_low":105.7,"eci_rank":216,"key":"llama-2-70b","name":"Llama 2-70B","org":"Meta AI","org_first":"Meta AI","released":"2023-07-18","scores":{"gpqa-diamond":26.3,"gsm8k":69.6,"hellaswag":85.3,"math-level-5":3.3,"mmlu":69.9,"winogrande":80.2},"weights":"open"},{"access":"Unreleased","benchmarks":3,"country":"United States of America","eci":105.2,"eci_high":110.2,"eci_low":94.5,"eci_rank":232,"key":"llama-2-34b","name":"Llama 2-34B","org":"Meta AI","org_first":"Meta AI","released":"2023-07-18","scores":{"gsm8k":42.2,"mmlu":62.6,"winogrande":76.7},"weights":"closed"},{"access":"Open weights (restricted use)","benchmarks":4,"country":"United States of America","eci":106.2,"eci_high":110.8,"eci_low":96.1,"eci_rank":231,"key":"llama-2-13b","name":"Llama 2-13B","org":"Meta AI","org_first":"Meta AI","released":"2023-07-18","scores":{"gsm8k":36.9,"hellaswag":80.7,"mmlu":55.6,"winogrande":72.8},"weights":"open"},{"access":"API access","benchmarks":3,"country":"United States of America","eci":120.1,"eci_high":123.8,"eci_low":113,"eci_rank":197,"key":"claude-2","name":"Claude 2","org":"Anthropic","org_first":"Anthropic","released":"2023-07-11","scores":{"gpqa-diamond":34.7,"math-level-5":11.7,"mmlu":78.5},"weights":"closed"},{"access":"Open weights (restricted use)","benchmarks":2,"country":"China","key":"baichuan-1-13b","name":"Baichuan 1-13B","org":"Baichuan","org_first":"Baichuan","released":"2023-07-11","scores":{"gsm8k":26.8,"mmlu":51.6},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":3,"country":"United States of America","eci":93.4,"eci_high":100.8,"eci_low":78.8,"eci_rank":250,"key":"xgen-7b","name":"XGen-7B","org":"Salesforce","org_first":"Salesforce","released":"2023-06-27","scores":{"hellaswag":74.2,"mmlu":36.3,"winogrande":64.9},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":3,"country":"United States of America","eci":100.7,"eci_high":106.4,"eci_low":88.9,"eci_rank":240,"key":"mpt-30b","name":"MPT-30B","org":"MosaicML","org_first":"MosaicML","released":"2023-06-22","scores":{"gsm8k":34.4,"mmlu":47.9,"winogrande":71},"weights":"open"},{"access":"Hosted access (no API)","benchmarks":1,"country":"United States of America","key":"inflection-1","name":"Inflection-1","org":"Inflection AI","org_first":"Inflection AI","released":"2023-06-22","scores":{"mmlu":72.7},"weights":"closed"},{"access":"Open weights (restricted use)","benchmarks":1,"country":"United States of America","key":"vicuna-13b-v1-3","name":"Vicuna-13B-v1.3","org":"Large Model Systems Organization, University of California (UC) Berkeley","org_first":"Large Model Systems Organization","released":"2023-06-18","scores":{"gsm8k":22.6},"weights":"open"},{"access":"API access","benchmarks":4,"country":"United States of America","eci":123.1,"eci_high":126.5,"eci_low":116.6,"eci_rank":187,"key":"gpt-4-jun-2023","name":"GPT-4 (Jun 2023)","org":"OpenAI","org_first":"OpenAI","released":"2023-06-13","scores":{"gpqa-diamond":30.7,"gsm8k":90,"math-level-5":23,"mmlu":82.4},"weights":"closed"},{"access":"API access","benchmarks":3,"country":"United States of America","eci":113.4,"eci_high":117.8,"eci_low":104.3,"eci_rank":217,"key":"gpt-3-5-turbo-jun-2023","name":"GPT-3.5 Turbo (Jun 2023)","org":"OpenAI","org_first":"OpenAI","released":"2023-06-13","scores":{"gsm8k":57.8,"mmlu":68.9,"winogrande":81.6},"weights":"closed"},{"access":"Open weights (non-commercial)","benchmarks":2,"country":"China","eci":90.5,"eci_high":98.5,"eci_low":75.6,"eci_rank":252,"key":"baichuan1-7b","name":"Baichuan1-7B","org":"Baichuan","org_first":"Baichuan","released":"2023-06-01","scores":{"gsm8k":9.2,"mmlu":42.3},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":4,"country":"United States of America","eci":94.6,"eci_high":101.6,"eci_low":80.5,"eci_rank":247,"key":"mpt-7b","name":"MPT-7B","org":"MosaicML","org_first":"MosaicML","released":"2023-05-05","scores":{"gsm8k":9.1,"hellaswag":76.4,"mmlu":30.8,"winogrande":68.6},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":4,"country":"United Arab Emirates","eci":95.1,"eci_high":102.4,"eci_low":81.1,"eci_rank":246,"key":"falcon-7b","name":"Falcon-7B","org":"Technology Innovation Institute","org_first":"Technology Innovation Institute","released":"2023-04-24","scores":{"gsm8k":6.8,"hellaswag":78.1,"mmlu":35,"winogrande":67.2},"weights":"open"},{"access":"API access","benchmarks":1,"country":"United States of America","key":"claude-1-3","name":"Claude 1.3","org":"Anthropic","org_first":"Anthropic","released":"2023-04-18","scores":{"mmlu":77},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":3,"country":"United States of America","eci":89.7,"eci_high":97.5,"eci_low":73.6,"eci_rank":253,"key":"dolly-2-0-12b","name":"Dolly 2.0-12b","org":"Databricks","org_first":"Databricks","released":"2023-04-11","scores":{"hellaswag":70.8,"mmlu":26.2,"winogrande":61.8},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":3,"country":"United States of America","eci":83.2,"eci_high":92,"eci_low":66,"eci_rank":256,"key":"cerebras-gpt-13b","name":"Cerebras-GPT-13B","org":"Cerebras Systems","org_first":"Cerebras Systems","released":"2023-03-20","scores":{"hellaswag":59.4,"mmlu":26.2,"winogrande":60.8},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":4,"country":"United Arab Emirates","eci":104.5,"eci_high":109.9,"eci_low":92.8,"eci_rank":235,"key":"falcon-40b","name":"Falcon-40B","org":"Technology Innovation Institute","org_first":"Technology Innovation Institute","released":"2023-03-15","scores":{"gsm8k":33.8,"hellaswag":85.3,"mmlu":56.9,"winogrande":76.9},"weights":"open"},{"access":"API access","benchmarks":6,"country":"United States of America","eci":125.9,"eci_high":130.4,"eci_low":119.2,"eci_rank":180,"key":"gpt-4-mar-2023","name":"GPT-4 (Mar 2023)","org":"OpenAI","org_first":"OpenAI","released":"2023-03-14","scores":{"gpqa-diamond":35.7,"gsm8k":92,"hellaswag":95.3,"metr-time-horizon":5.36,"mmlu":86.4,"winogrande":87.5},"weights":"closed"},{"access":"Open weights (non-commercial)","benchmarks":4,"country":"United States of America","eci":96.7,"eci_high":103.7,"eci_low":83.5,"eci_rank":243,"key":"llama-7b","name":"LLaMA-7B","org":"Meta AI","org_first":"Meta AI","released":"2023-02-24","scores":{"gsm8k":11,"hellaswag":76.2,"mmlu":35.6,"winogrande":70.1},"weights":"open"},{"access":"Open weights (non-commercial)","benchmarks":4,"country":"United States of America","eci":110.2,"eci_high":114.1,"eci_low":101.4,"eci_rank":224,"key":"llama-65b","name":"LLaMA-65B","org":"Meta AI","org_first":"Meta AI","released":"2023-02-24","scores":{"gsm8k":54.4,"hellaswag":84.2,"mmlu":63.4,"winogrande":77},"weights":"open"},{"access":"Open weights (non-commercial)","benchmarks":4,"country":"United States of America","eci":107.4,"eci_high":112.1,"eci_low":97.4,"eci_rank":229,"key":"llama-33b","name":"LLaMA-33B","org":"Meta AI","org_first":"Meta AI","released":"2023-02-24","scores":{"gsm8k":44.1,"hellaswag":82.8,"mmlu":58.7,"winogrande":76},"weights":"open"},{"access":"Open weights (non-commercial)","benchmarks":4,"country":"United States of America","eci":100.6,"eci_high":106.4,"eci_low":89,"eci_rank":241,"key":"llama-13b","name":"LLaMA-13B","org":"Meta AI","org_first":"Meta AI","released":"2023-02-24","scores":{"gsm8k":20.5,"hellaswag":79.2,"mmlu":47.7,"winogrande":73},"weights":"open"},{"access":"","benchmarks":2,"country":"United States of America","key":"davinci-003","name":"davinci-003","org":"OpenAI","org_first":"OpenAI","released":"2022-11-28","scores":{"gsm8k":78.2,"hellaswag":85.5},"weights":"unknown"},{"access":"Open weights (restricted use)","benchmarks":2,"country":"France,United States of America","key":"bloom-176b","name":"BLOOM-176B","org":"Hugging Face, BigScience","org_first":"Hugging Face","released":"2022-07-06","scores":{"gsm8k":9.5,"hellaswag":74.4},"weights":"open"},{"access":"Open weights (non-commercial)","benchmarks":2,"country":"United States of America","key":"opt-1-3b","name":"OPT-1.3B","org":"Meta AI","org_first":"Meta AI","released":"2022-05-11","scores":{"hellaswag":41.5,"winogrande":61},"weights":"open"},{"access":"Open weights (non-commercial)","benchmarks":2,"country":"United States of America","key":"opt-66b","name":"OPT-66B","org":"Meta AI","org_first":"Meta AI","released":"2022-05-03","scores":{"gsm8k":1.8,"hellaswag":74.5},"weights":"open"},{"access":"Open weights (non-commercial)","benchmarks":2,"country":"United States of America","key":"opt-175b","name":"OPT-175B","org":"Meta AI","org_first":"Meta AI","released":"2022-05-02","scores":{"gsm8k":4,"hellaswag":79.1},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":2,"country":"United States of America","key":"gpt-neox-20b","name":"GPT-NeoX-20B","org":"EleutherAI","org_first":"EleutherAI","released":"2022-04-07","scores":{"hellaswag":70.5,"winogrande":66.1},"weights":"open"},{"access":"Unreleased","benchmarks":4,"country":"United States of America","key":"palm-540b","name":"PaLM (540B)","org":"Google Research","org_first":"Google Research","released":"2022-04-04","scores":{"gsm8k":56.5,"hellaswag":83.8,"mmlu":69.3,"winogrande":85.1},"weights":"closed"},{"access":"Unreleased","benchmarks":2,"country":"United Kingdom","key":"chinchilla","name":"Chinchilla","org":"DeepMind","org_first":"DeepMind","released":"2022-03-29","scores":{"mmlu":67.5,"winogrande":74.9},"weights":"closed"},{"access":"API access","benchmarks":4,"country":"United States of America","key":"davinci-002","name":"davinci-002","org":"OpenAI","org_first":"OpenAI","released":"2022-03-15","scores":{"gsm8k":56.8,"hellaswag":81.5,"mmlu":70,"winogrande":81.6},"weights":"closed"},{"access":"API access","benchmarks":2,"country":"United States of America","key":"instructgpt-6b","name":"InstructGPT 6B","org":"OpenAI","org_first":"OpenAI","released":"2022-01-27","scores":{"gsm8k":0.6,"hellaswag":67.6},"weights":"closed"},{"access":"","benchmarks":2,"country":"United States of America","key":"instructgpt-350m","name":"InstructGPT 350M","org":"OpenAI","org_first":"OpenAI","released":"2022-01-27","scores":{"gsm8k":0.4,"hellaswag":42.9},"weights":"unknown"},{"access":"API access","benchmarks":3,"country":"United States of America","key":"instructgpt-175b","name":"InstructGPT 175B","org":"OpenAI","org_first":"OpenAI","released":"2022-01-27","scores":{"hellaswag":79.3,"mmlu":43.9,"winogrande":77.7},"weights":"closed"},{"access":"API access","benchmarks":2,"country":"United States of America","key":"instructgpt-1-3b","name":"InstructGPT 1.3B","org":"OpenAI","org_first":"OpenAI","released":"2022-01-27","scores":{"gsm8k":0,"hellaswag":56.1},"weights":"closed"},{"access":"Unreleased","benchmarks":2,"country":"United States of America","key":"glam","name":"GLaM","org":"Google","org_first":"Google","released":"2021-12-13","scores":{"hellaswag":77.2,"winogrande":79.2},"weights":"closed"},{"access":"Unreleased","benchmarks":3,"country":"United Kingdom","key":"gopher-280b","name":"Gopher (280B)","org":"DeepMind","org_first":"DeepMind","released":"2021-12-08","scores":{"hellaswag":79.2,"mmlu":60,"winogrande":70.2},"weights":"closed"},{"access":"Unreleased","benchmarks":2,"country":"United States of America","key":"megatron-turing-nlg-530b","name":"Megatron-Turing NLG 530B","org":"Microsoft, Nvidia","org_first":"Microsoft","released":"2021-10-11","scores":{"hellaswag":82.4,"winogrande":78.9},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":3,"country":"United States of America,Germany","key":"gpt-j-6b","name":"GPT-J-6B","org":"EleutherAI, LAION","org_first":"EleutherAI","released":"2021-08-05","scores":{"hellaswag":66.2,"mmlu":25.7,"winogrande":64.5},"weights":"open"},{"access":"Open weights (unrestricted)","benchmarks":1,"country":"United States of America","key":"gpt-neo-2-7b","name":"GPT-Neo-2.7B","org":"EleutherAI","org_first":"EleutherAI","released":"2021-03-30","scores":{"hellaswag":42.7},"weights":"open"},{"access":"","benchmarks":2,"country":"United States of America","key":"gpt-3-xl-babbage","name":"GPT-3 XL (babbage)","org":"OpenAI","org_first":"OpenAI","released":"2020-06-22","scores":{"gsm8k":0.7,"hellaswag":55.5},"weights":"unknown"},{"access":"","benchmarks":2,"country":"United States of America","key":"gpt-3-medium","name":"GPT-3 Medium","org":"OpenAI","org_first":"OpenAI","released":"2020-06-22","scores":{"gsm8k":0.6,"hellaswag":43.5},"weights":"unknown"},{"access":"","benchmarks":2,"country":"United States of America","key":"gpt-3-6-7b","name":"GPT-3 6.7B","org":"OpenAI","org_first":"OpenAI","released":"2020-06-22","scores":{"gsm8k":1.6,"hellaswag":68.2},"weights":"unknown"},{"access":"API access","benchmarks":3,"country":"United States of America","key":"gpt-3-175b-davinci","name":"GPT-3 175B (davinci)","org":"OpenAI","org_first":"OpenAI","released":"2020-05-28","scores":{"gsm8k":9,"hellaswag":77.5,"metr-time-horizon":0.14},"weights":"closed"},{"access":"Open weights (unrestricted)","benchmarks":3,"country":"United States of America","key":"gpt-2-1-5b","name":"GPT-2 (1.5B)","org":"OpenAI","org_first":"OpenAI","released":"2019-11-05","scores":{"hellaswag":40,"metr-time-horizon":0.05,"winogrande":58.3},"weights":"open"}],"retrieved":"2026-10-12","source":"Epoch AI, 'Capabilities \u0026 benchmarking', epoch.ai/benchmarks"},"modified":"2026-10-12","pages":[{"description":"Which language model is the most capable overall? 257 LLMs ranked by the Epoch Capabilities Index, plus the best open-weights models, the best model of each organization and the leaders on 15 benchmarks.","json":"https://abxk.ai/en/llm/ranking/index.json","kind":"page","title":"LLM Ranking: 257 Language Models on One Scale","url":"https://abxk.ai/en/llm/ranking/"},{"description":"Which model held the record when? 16 LLM benchmarks from GPT-2 to today, with every record by release date and today's leaders — built from Epoch AI's open benchmark archive.","json":"https://abxk.ai/en/llm/benchmarks/index.json","kind":"page","title":"LLM Benchmarks 2019–2026: The Records on 16 Leaderboards","url":"https://abxk.ai/en/llm/benchmarks/"}],"section":"llm","seo_title":"LLM Database: 340 Language Models Compared | ABXK","title":"LLM Database","translations":[{"json":"https://abxk.ai/de/llm/index.json","language":"de","url":"https://abxk.ai/de/llm/"}],"url":"https://abxk.ai/en/llm/"}