{"data":{"sources":[{"source":"epoch","name":"Epoch AI Benchmarking Hub","last_ok":"2026-10-31T00:55:45+00:00","last_error":"","rows":6842,"newest_result":"2026-10-28","stale":false,"reasons":[],"page":"https://epoch.ai/benchmarks","licence":"CC BY 4.0, credit Epoch AI"},{"source":"lmarena","name":"LMArena","last_ok":"2026-10-31T00:59:47+00:00","last_error":"","rows":738,"newest_result":"2026-10-28","stale":false,"reasons":[],"page":"https://lmarena.ai/leaderboard","licence":"CC BY 4.0 dataset (lmarena-ai/leaderboard-dataset on Hugging Face)"},{"source":"self_reported","name":"Lab-claimed (blogs and model cards)","last_ok":"2026-10-30T06:53:44+00:00","last_error":"","rows":1,"newest_result":"2026-10-30","stale":false,"reasons":[],"page":"","licence":"claims by the lab, unverified; extracted by one capped model call a day"},{"source":"swebench","name":"SWE-bench leaderboards","last_ok":"2026-10-31T00:55:46+00:00","last_error":"","rows":323,"newest_result":"2026-02-26","stale":true,"reasons":["not updating upstream (newest result 247 days old)"],"page":"https://www.swebench.com","licence":"public GitHub data; credit SWE-bench"}],"labs":[{"id":"openai","name":"OpenAI","models":405,"records":340},{"id":"anthropic","name":"Anthropic","models":343,"records":242},{"id":"google","name":"Google","models":216,"records":105},{"id":"unknown","name":"Unknown","models":232,"records":29},{"id":"meta","name":"Meta","models":129,"records":14},{"id":"deepseek","name":"DeepSeek","models":116,"records":12},{"id":"xai","name":"xAI","models":82,"records":10},{"id":"eleutherai","name":"EleutherAI","models":3,"records":9},{"id":"microsoft","name":"Microsoft","models":35,"records":8},{"id":"aws","name":"AWS","models":6,"records":8},{"id":"alibaba","name":"Alibaba","models":208,"records":7},{"id":"moonshot","name":"Moonshot","models":43,"records":4},{"id":"technology-innovation-institute","name":"Technology Innovation Institute","models":9,"records":4},{"id":"mistral","name":"Mistral","models":94,"records":3},{"id":"amazon","name":"Amazon","models":28,"records":2},{"id":"bytedance","name":"ByteDance","models":7,"records":2},{"id":"inflection-ai","name":"Inflection AI","models":1,"records":2},{"id":"openhands","name":"OpenHands","models":2,"records":1},{"id":"stability-ai","name":"Stability AI","models":2,"records":1},{"id":"atlassian","name":"Atlassian","models":1,"records":1},{"id":"cursor","name":"Cursor","models":1,"records":1},{"id":"salesforce-research","name":"Salesforce Research","models":1,"records":1},{"id":"z-ai","name":"Z.ai","models":38,"records":0},{"id":"zai","name":"zai","models":33,"records":0},{"id":"nvidia","name":"NVIDIA","models":30,"records":0}],"benchmarks":[{"id":"lmarena-text","name":"LMArena text","source":"lmarena","unit":"elo","ceiling":null,"models":809,"as_of":"2026-10-28","best":{"d":"2026-10-28","s":1540.04,"m":"Gemini 4 Carbon (High)","l":"google","weights":"closed"},"saturation":null,"change_12m":null,"open_lag_days":13,"gap_closed_minus_open":48.69,"new_records":[{"d":"2026-10-15","s":1532.34,"m":"Gemini 4 Argon (High)","l":"google"},{"d":"2026-10-28","s":1540.04,"m":"Gemini 4 Carbon (High)","l":"google"}],"claimed":0},{"id":"epoch-gpqa-diamond","name":"GPQA diamond","source":"epoch","unit":"fraction","ceiling":0.97,"models":326,"as_of":"2026-10-26","best":{"d":"2026-09-30","s":0.964015,"m":"Gemini 4 Argon (high)","l":"google","weights":"closed"},"saturation":0.9938,"change_12m":0.094015,"open_lag_days":240,"gap_closed_minus_open":0.032197,"new_records":[{"d":"2026-09-30","s":0.964015,"m":"Gemini 4 Argon (high)","l":"google"}],"claimed":0},{"id":"epoch-otis-mock-aime-2024-2025","name":"OTIS Mock AIME 2024-2025","source":"epoch","unit":"fraction","ceiling":1,"models":306,"as_of":"2026-10-26","best":{"d":"2026-04-23","s":1,"m":"GPT-5.5 Pro (xhigh)","l":"openai","weights":"closed"},"saturation":1,"change_12m":0.086111,"open_lag_days":177,"gap_closed_minus_open":0.005556,"new_records":[],"claimed":0},{"id":"epoch-eci","name":"Epoch Capabilities Index","source":"epoch","unit":"points","ceiling":null,"models":282,"as_of":"2026-10-26","best":{"d":"2026-09-03","s":169.06,"m":"GPT-6 Astra","l":"openai","weights":"closed"},"saturation":null,"change_12m":18.32,"open_lag_days":226,"gap_closed_minus_open":10.23,"new_records":[],"claimed":0},{"id":"epoch-arc-agi","name":"ARC-AGI","source":"epoch","unit":"fraction","ceiling":1,"models":278,"as_of":"2026-10-17","best":{"d":"2026-06-09","s":0.985,"m":"Claude Fable 5 (xhigh)","l":"anthropic","weights":"closed"},"saturation":0.985,"change_12m":0.283,"open_lag_days":147,"gap_closed_minus_open":0.04,"new_records":[],"claimed":0},{"id":"epoch-arc-agi-2","name":"ARC-AGI-2","source":"epoch","unit":"fraction","ceiling":1,"models":266,"as_of":"2026-10-17","best":{"d":"2026-09-03","s":0.95,"m":"GPT-6 Astra (max)","l":"openai","weights":"closed"},"saturation":0.95,"change_12m":0.7667,"open_lag_days":226,"gap_closed_minus_open":0.169444,"new_records":[],"claimed":0},{"id":"epoch-critpt","name":"CritPt","source":"epoch","unit":"fraction","ceiling":1,"models":256,"as_of":"2026-10-28","best":{"d":"2026-10-28","s":0.36,"m":"Gemini 4 Carbon (high)","l":"google","weights":"closed"},"saturation":0.36,"change_12m":0.234,"open_lag_days":247,"gap_closed_minus_open":0.12,"new_records":[{"d":"2026-10-28","s":0.36,"m":"Gemini 4 Carbon (high)","l":"google"}],"claimed":0},{"id":"epoch-dtbench","name":"DTBench","source":"epoch","unit":"fraction","ceiling":1,"models":233,"as_of":"2026-10-12","best":{"d":"2026-09-22","s":0.9893,"m":"Claude Opus 5.5 (max)","l":"anthropic","weights":"closed"},"saturation":0.9893,"change_12m":0.0826,"open_lag_days":175,"gap_closed_minus_open":0.0506,"new_records":[{"d":"2026-09-22","s":0.9893,"m":"Claude Opus 5.5 (max)","l":"anthropic"}],"claimed":0},{"id":"epoch-scicode","name":"SciCode","source":"epoch","unit":"fraction","ceiling":1,"models":226,"as_of":"2026-10-28","best":{"d":"2026-09-22","s":0.668981,"m":"Claude Opus 5.5 (max)","l":"anthropic","weights":"closed"},"saturation":0.669,"change_12m":0.206018,"open_lag_days":37,"gap_closed_minus_open":0.074074,"new_records":[{"d":"2026-09-22","s":0.668981,"m":"Claude Opus 5.5 (max)","l":"anthropic"}],"claimed":0},{"id":"epoch-lmca","name":"LMCA","source":"epoch","unit":"fraction","ceiling":0.85,"models":195,"as_of":"2026-10-12","best":{"d":"2026-09-28","s":0.6826,"m":"Claude Sonnet 5.5 (max)","l":"anthropic","weights":"closed"},"saturation":0.8031,"change_12m":0.2827,"open_lag_days":190,"gap_closed_minus_open":0.1275,"new_records":[{"d":"2026-09-22","s":0.6823,"m":"Claude Opus 5.5 (max)","l":"anthropic"},{"d":"2026-09-28","s":0.6826,"m":"Claude Sonnet 5.5 (max)","l":"anthropic"}],"claimed":0},{"id":"swebench-verified","name":"SWE-bench Verified (agent submissions)","source":"swebench","unit":"fraction","ceiling":1,"models":180,"as_of":"2026-02-26","best":{"d":"2025-12-05","s":0.792,"m":"Claude 4.5 Opus","l":"anthropic","weights":"closed"},"saturation":0.792,"change_12m":0.004,"open_lag_days":197,"gap_closed_minus_open":0.034,"new_records":[],"claimed":0},{"id":"epoch-weirdml","name":"WeirdML","source":"epoch","unit":"fraction","ceiling":1,"models":170,"as_of":"2026-09-03","best":{"d":"2026-09-03","s":0.9357,"m":"GPT-6 Astra (pro, max)","l":"openai","weights":"closed"},"saturation":0.9357,"change_12m":0.3287,"open_lag_days":84,"gap_closed_minus_open":0.11,"new_records":[],"claimed":0},{"id":"lmarena-vision","name":"LMArena vision","source":"lmarena","unit":"elo","ceiling":null,"models":167,"as_of":"2026-10-19","best":{"d":"2026-10-19","s":1324.88,"m":"Claude Fable 5 (High)","l":"anthropic","weights":"closed"},"saturation":null,"change_12m":null,"open_lag_days":0,"gap_closed_minus_open":31.01,"new_records":[{"d":"2026-10-19","s":1324.88,"m":"Claude Fable 5 (High)","l":"anthropic"}],"claimed":0},{"id":"epoch-chess-puzzles","name":"Chess Puzzles","source":"epoch","unit":"fraction","ceiling":1,"models":154,"as_of":"2026-10-17","best":{"d":"2026-09-03","s":0.76,"m":"GPT-6 Astra (pro, max)","l":"openai","weights":"closed"},"saturation":0.76,"change_12m":0.28,"open_lag_days":337,"gap_closed_minus_open":0.28,"new_records":[],"claimed":0},{"id":"lmarena-webdev","name":"LMArena web development","source":"lmarena","unit":"elo","ceiling":null,"models":153,"as_of":"2026-10-28","best":{"d":"2026-10-28","s":1809.84,"m":"Claude Opus 5.5 (Max)","l":"anthropic","weights":"closed"},"saturation":null,"change_12m":null,"open_lag_days":11,"gap_closed_minus_open":143.58,"new_records":[{"d":"2026-10-16","s":1807.31,"m":"Claude Opus 5.5 (Max)","l":"anthropic"},{"d":"2026-10-27","s":1808.74,"m":"Claude Opus 5.5 (Max)","l":"anthropic"},{"d":"2026-10-28","s":1809.84,"m":"Claude Opus 5.5 (Max)","l":"anthropic"}],"claimed":0},{"id":"epoch-mystery-game-puzzles","name":"Mystery Game Puzzles","source":"epoch","unit":"fraction","ceiling":1,"models":142,"as_of":"2026-10-26","best":{"d":"2026-09-03","s":0.84,"m":"GPT-6 Astra (max)","l":"openai","weights":"closed"},"saturation":0.84,"change_12m":0.55,"open_lag_days":177,"gap_closed_minus_open":0.35,"new_records":[],"claimed":0},{"id":"epoch-mmlu","name":"MMLU","source":"epoch","unit":"fraction","ceiling":1,"models":135,"as_of":"2024-12-26","best":{"d":"2024-11-20","s":0.881,"m":"GPT-4o (Nov 2024)","l":"openai","weights":"closed"},"saturation":0.881,"change_12m":0,"open_lag_days":65,"gap_closed_minus_open":0.01,"new_records":[],"claimed":0},{"id":"epoch-webdev-arena","name":"WebDev Arena","source":"epoch","unit":"points","ceiling":null,"models":129,"as_of":"2026-09-30","best":{"d":"2026-09-22","s":1820.31,"m":"Claude Opus 5.5 (max)","l":"anthropic","weights":"closed"},"saturation":null,"change_12m":400.79,"open_lag_days":-8,"gap_closed_minus_open":146.05,"new_records":[{"d":"2026-09-22","s":1820.31,"m":"Claude Opus 5.5 (max)","l":"anthropic"}],"claimed":0},{"id":"epoch-ale-bench","name":"ALE-Bench","source":"epoch","unit":"points","ceiling":null,"models":127,"as_of":"2026-10-26","best":{"d":"2026-09-03","s":2951.32,"m":"GPT-6 Astra (max)","l":"openai","weights":"closed"},"saturation":null,"change_12m":1788.87,"open_lag_days":254,"gap_closed_minus_open":1329.32,"new_records":[],"claimed":0},{"id":"epoch-frontiermath-tiers-1-3-v2-private","name":"FrontierMath-Tiers-1-3-v2-Private","source":"epoch","unit":"fraction","ceiling":0.95,"models":122,"as_of":"2026-10-26","best":{"d":"2026-09-03","s":0.936842,"m":"GPT-6 Astra (max)","l":"openai","weights":"closed"},"saturation":0.9861,"change_12m":0.378947,"open_lag_days":217,"gap_closed_minus_open":0.215011,"new_records":[],"claimed":0},{"id":"epoch-math-level-5","name":"MATH level 5","source":"epoch","unit":"fraction","ceiling":1,"models":108,"as_of":"2025-10-15","best":{"d":"2025-08-07","s":0.981307,"m":"GPT-5 (high)","l":"openai","weights":"closed"},"saturation":0.9813,"change_12m":0,"open_lag_days":42,"gap_closed_minus_open":0.014917,"new_records":[],"claimed":0},{"id":"epoch-simplebench","name":"SimpleBench","source":"epoch","unit":"fraction","ceiling":1,"models":102,"as_of":"2026-08-12","best":{"d":"2026-06-09","s":0.819,"m":"Claude Fable 5 (max)","l":"anthropic","weights":"closed"},"saturation":0.819,"change_12m":0.195,"open_lag_days":421,"gap_closed_minus_open":0.208,"new_records":[],"claimed":0},{"id":"epoch-frontiermath-2025-02-28-private","name":"FrontierMath-2025-02-28-Private","source":"epoch","unit":"fraction","ceiling":0.57,"models":101,"as_of":"2026-05-28","best":{"d":"2026-04-23","s":0.524,"m":"GPT-5.5 Pro (high)","l":"openai","weights":"closed"},"saturation":0.9193,"change_12m":0.199862,"open_lag_days":130,"gap_closed_minus_open":0.134345,"new_records":[],"claimed":0},{"id":"epoch-forecastbench","name":"ForecastBench","source":"epoch","unit":"points","ceiling":null,"models":97,"as_of":"2026-08-13","best":{"d":"2026-04-23","s":67,"m":"GPT-5.5 (high)","l":"openai","weights":"closed"},"saturation":null,"change_12m":1.7,"open_lag_days":504,"gap_closed_minus_open":2.3,"new_records":[],"claimed":0},{"id":"epoch-simpleqa-verified","name":"SimpleQA Verified","source":"epoch","unit":"fraction","ceiling":1,"models":93,"as_of":"2026-10-26","best":{"d":"2026-09-30","s":0.76,"m":"Gemini 4 Argon (high)","l":"google","weights":"closed"},"saturation":0.76,"change_12m":0.259,"open_lag_days":302,"gap_closed_minus_open":0.194,"new_records":[{"d":"2026-09-30","s":0.76,"m":"Gemini 4 Argon (high)","l":"google"}],"claimed":0},{"id":"epoch-gsm8k","name":"GSM8K","source":"epoch","unit":"fraction","ceiling":1,"models":85,"as_of":"2024-11-29","best":{"d":"2024-06-17","s":0.945,"m":"DeepSeek-Coder-V2 236B","l":"deepseek","weights":"open"},"saturation":0.945,"change_12m":0,"open_lag_days":null,"gap_closed_minus_open":-0.025,"new_records":[],"claimed":0},{"id":"swebench-lite","name":"SWE-bench Lite (agent submissions)","source":"swebench","unit":"fraction","ceiling":1,"models":84,"as_of":"2025-09-11","best":{"d":"2025-06-25","s":0.6033,"m":"Claude 4 Sonnet","l":"anthropic","weights":"closed"},"saturation":0.6033,"change_12m":0,"open_lag_days":230,"gap_closed_minus_open":0.1066,"new_records":[],"claimed":0},{"id":"epoch-cursorbench","name":"CursorBench","source":"epoch","unit":"fraction","ceiling":1,"models":82,"as_of":"2026-10-17","best":{"d":"2026-09-22","s":0.578,"m":"Claude Opus 5.5 (max)","l":"anthropic","weights":"closed"},"saturation":0.578,"change_12m":null,"open_lag_days":46,"gap_closed_minus_open":0.101,"new_records":[{"d":"2026-09-22","s":0.578,"m":"Claude Opus 5.5 (max)","l":"anthropic"}],"claimed":0},{"id":"epoch-proofbench","name":"ProofBench","source":"epoch","unit":"fraction","ceiling":1,"models":82,"as_of":"2026-10-23","best":{"d":"2026-09-01","s":1,"m":"Claude Fable 5.1 (max)","l":"anthropic","weights":"closed"},"saturation":1,"change_12m":0.81,"open_lag_days":37,"gap_closed_minus_open":0.13,"new_records":[],"claimed":0},{"id":"epoch-winogrande","name":"Winogrande","source":"epoch","unit":"fraction","ceiling":1,"models":80,"as_of":"2024-12-26","best":{"d":"2024-07-23","s":0.892,"m":"Llama 3.1-405B","l":"meta","weights":"open"},"saturation":0.892,"change_12m":0,"open_lag_days":null,"gap_closed_minus_open":-0.007,"new_records":[],"claimed":0},{"id":"epoch-frontiermath-tier-4-v2-private","name":"FrontierMath-Tier-4-v2-Private","source":"epoch","unit":"fraction","ceiling":1,"models":78,"as_of":"2026-10-26","best":{"d":"2026-09-29","s":1,"m":"GPT-6.1 Sol (max)","l":"openai","weights":"closed"},"saturation":1,"change_12m":0.780488,"open_lag_days":226,"gap_closed_minus_open":0.536585,"new_records":[{"d":"2026-09-29","s":1,"m":"GPT-6.1 Sol (max)","l":"openai"}],"claimed":0},{"id":"epoch-arc-ai2","name":"ARC AI2","source":"epoch","unit":"fraction","ceiling":1,"models":76,"as_of":"2024-12-26","best":{"d":"2024-07-23","s":0.953,"m":"Llama 3.1-405B","l":"meta","weights":"open"},"saturation":0.953,"change_12m":0,"open_lag_days":null,"gap_closed_minus_open":-0.079,"new_records":[],"claimed":0},{"id":"epoch-frontiermath-tier-4-2025-07-01-private","name":"FrontierMath-Tier-4-2025-07-01-Private","source":"epoch","unit":"fraction","ceiling":0.6,"models":72,"as_of":"2026-05-28","best":{"d":"2026-05-08","s":0.479,"m":"AI co-mathematician","l":"google","weights":"closed"},"saturation":0.7983,"change_12m":0.333,"open_lag_days":195,"gap_closed_minus_open":0.3332,"new_records":[],"claimed":0},{"id":"epoch-aider-polyglot","name":"Aider polyglot","source":"epoch","unit":"fraction","ceiling":1,"models":71,"as_of":"2025-12-01","best":{"d":"2025-08-07","s":0.88,"m":"GPT-5 (high)","l":"openai","weights":"closed"},"saturation":0.88,"change_12m":0,"open_lag_days":166,"gap_closed_minus_open":0.138,"new_records":[],"claimed":0},{"id":"epoch-vending-bench-2","name":"Vending-Bench 2","source":"epoch","unit":"points","ceiling":null,"models":71,"as_of":"2026-10-28","best":{"d":"2026-10-28","s":16484.206667,"m":"Gemini 4 Carbon (unknown)","l":"google","weights":"closed"},"saturation":null,"change_12m":12645.466667,"open_lag_days":61,"gap_closed_minus_open":8170.428334,"new_records":[{"d":"2026-10-28","s":16484.206667,"m":"Gemini 4 Carbon (unknown)","l":"google"}],"claimed":0},{"id":"epoch-hellaswag","name":"HellaSwag","source":"epoch","unit":"fraction","ceiling":1,"models":70,"as_of":"2024-12-26","best":{"d":"2023-03-14","s":0.953,"m":"GPT-4 (Mar 2023)","l":"openai","weights":"closed"},"saturation":0.953,"change_12m":0,"open_lag_days":497,"gap_closed_minus_open":0.061,"new_records":[],"claimed":0},{"id":"epoch-deepswe","name":"DeepSWE","source":"epoch","unit":"fraction","ceiling":1,"models":69,"as_of":"2026-09-03","best":{"d":"2026-09-03","s":0.74115,"m":"GPT-6 Astra (xhigh)","l":"openai","weights":"closed"},"saturation":0.7411,"change_12m":null,"open_lag_days":66,"gap_closed_minus_open":0.051571,"new_records":[],"claimed":0},{"id":"epoch-boolq","name":"BoolQ","source":"epoch","unit":"fraction","ceiling":1,"models":65,"as_of":"2024-08-17","best":{"d":"2023-05-17","s":0.909,"m":"PaLM 2-L","l":"unknown","weights":"unknown"},"saturation":0.909,"change_12m":0,"open_lag_days":28,"gap_closed_minus_open":0.003,"new_records":[],"claimed":0},{"id":"epoch-fiction-livebench","name":"Fiction.LiveBench","source":"epoch","unit":"fraction","ceiling":1,"models":61,"as_of":"2026-01-27","best":{"d":"2025-06-10","s":0.972,"m":"o3-pro","l":"openai","weights":"closed"},"saturation":0.972,"change_12m":0,"open_lag_days":286,"gap_closed_minus_open":0.111,"new_records":[],"claimed":0},{"id":"epoch-apex-agents","name":"APEX-Agents","source":"epoch","unit":"fraction","ceiling":1,"models":60,"as_of":"2026-10-26","best":{"d":"2026-09-30","s":0.824,"m":"Gemini 4 Argon (unknown)","l":"google","weights":"closed"},"saturation":0.824,"change_12m":0.78,"open_lag_days":25,"gap_closed_minus_open":0.126,"new_records":[{"d":"2026-09-22","s":0.735,"m":"Claude Opus 5.5 (max)","l":"anthropic"},{"d":"2026-09-28","s":0.755,"m":"Claude Sonnet 5.5 (max)","l":"anthropic"},{"d":"2026-09-30","s":0.824,"m":"Gemini 4 Argon (unknown)","l":"google"}],"claimed":0},{"id":"epoch-piqa","name":"PIQA","source":"epoch","unit":"fraction","ceiling":1,"models":59,"as_of":"2024-12-26","best":{"d":"2024-07-18","s":0.887,"m":"GPT-4o mini","l":"openai","weights":"closed"},"saturation":0.887,"change_12m":0,"open_lag_days":30,"gap_closed_minus_open":0.001,"new_records":[],"claimed":0},{"id":"epoch-terminal-bench","name":"Terminal Bench","source":"epoch","unit":"fraction","ceiling":1,"models":59,"as_of":"2026-04-23","best":{"d":"2026-03-05","s":0.818,"m":"GPT-5.4 (unknown thinking)","l":"openai","weights":"closed"},"saturation":0.818,"change_12m":0.405,"open_lag_days":85,"gap_closed_minus_open":0.293843,"new_records":[],"claimed":0},{"id":"epoch-gdp-pdf","name":"GDP.pdf","source":"epoch","unit":"fraction","ceiling":1,"models":56,"as_of":"2026-10-12","best":{"d":"2026-09-29","s":0.35,"m":"GPT-6.1 Sol (max)","l":"openai","weights":"closed"},"saturation":0.35,"change_12m":null,"open_lag_days":146,"gap_closed_minus_open":0.152,"new_records":[{"d":"2026-09-29","s":0.35,"m":"GPT-6.1 Sol (max)","l":"openai"}],"claimed":0},{"id":"epoch-hle","name":"HLE","source":"epoch","unit":"fraction","ceiling":1,"models":51,"as_of":"2026-10-07","best":{"d":"2026-09-03","s":0.548,"m":"GPT-6 Astra (unknown thinking)","l":"openai","weights":"closed"},"saturation":0.548,"change_12m":0.2316,"open_lag_days":173,"gap_closed_minus_open":0.3043,"new_records":[],"claimed":0},{"id":"epoch-bbh","name":"BBH","source":"epoch","unit":"fraction","ceiling":1,"models":50,"as_of":"2024-12-26","best":{"d":"2024-05-14","s":0.892,"m":"Gemini 1.5 Pro (May 2024)","l":"google","weights":"closed"},"saturation":0.892,"change_12m":0,"open_lag_days":226,"gap_closed_minus_open":0.017,"new_records":[],"claimed":0},{"id":"epoch-livebench","name":"LiveBench","source":"epoch","unit":"fraction","ceiling":1,"models":50,"as_of":"2025-11-13","best":{"d":"2025-03-25","s":0.8235,"m":"Gemini 2.5 Pro Exp (Mar 2025)","l":"google","weights":"closed"},"saturation":0.8235,"change_12m":0,"open_lag_days":78,"gap_closed_minus_open":0.1039,"new_records":[],"claimed":0},{"id":"epoch-videomme","name":"VideoMME","source":"epoch","unit":"fraction","ceiling":1,"models":50,"as_of":"2025-06-18","best":{"d":"2025-06-18","s":0.797,"m":"video-SALMONN 2+","l":"bytedance","weights":"unknown"},"saturation":0.797,"change_12m":0,"open_lag_days":217,"gap_closed_minus_open":0.015,"new_records":[],"claimed":0},{"id":"epoch-lech-mazur-writing","name":"Lech Mazur Writing","source":"epoch","unit":"fraction","ceiling":1,"models":49,"as_of":"2025-08-07","best":{"d":"2025-08-07","s":0.86,"m":"GPT-5 (medium)","l":"openai","weights":"closed"},"saturation":0.86,"change_12m":0,"open_lag_days":-26,"gap_closed_minus_open":0.004,"new_records":[],"claimed":0},{"id":"epoch-enigmaeval","name":"EnigmaEval","source":"epoch","unit":"fraction","ceiling":1,"models":46,"as_of":"2026-07-09","best":{"d":"2026-06-09","s":0.3928,"m":"Claude Fable 5 (high)","l":"anthropic","weights":"closed"},"saturation":0.3928,"change_12m":0.2053,"open_lag_days":406,"gap_closed_minus_open":0.359,"new_records":[],"claimed":0},{"id":"epoch-frontiercode","name":"FrontierCode","source":"epoch","unit":"fraction","ceiling":1,"models":46,"as_of":"2026-10-07","best":{"d":"2026-09-22","s":0.5464,"m":"Claude Opus 5.5 (medium)","l":"anthropic","weights":"closed"},"saturation":0.5464,"change_12m":null,"open_lag_days":49,"gap_closed_minus_open":0.1047,"new_records":[{"d":"2026-09-22","s":0.5464,"m":"Claude Opus 5.5 (medium)","l":"anthropic"}],"claimed":0},{"id":"epoch-metr-time-horizons","name":"METR Time Horizons","source":"epoch","unit":"minutes","ceiling":null,"models":45,"as_of":"2026-04-07","best":{"d":"2026-04-07","s":1044.780145,"m":"Claude Mythos Preview (Early)","l":"anthropic","weights":"closed"},"saturation":null,"change_12m":841.767568,"open_lag_days":255,"gap_closed_minus_open":990.585997,"new_records":[],"claimed":0},{"id":"epoch-gso-bench","name":"GSO-Bench","source":"epoch","unit":"fraction","ceiling":1,"models":42,"as_of":"2026-09-03","best":{"d":"2026-09-01","s":0.8824,"m":"Claude Fable 5.1 (unknown)","l":"anthropic","weights":"closed"},"saturation":0.8824,"change_12m":0.7353,"open_lag_days":87,"gap_closed_minus_open":0.8334,"new_records":[],"claimed":0},{"id":"lmarena-search","name":"LMArena search","source":"lmarena","unit":"elo","ceiling":null,"models":42,"as_of":"2026-10-13","best":{"d":"2026-10-13","s":1261.13,"m":"gpt-5.6-sol-xhigh","l":"openai","weights":"closed"},"saturation":null,"change_12m":null,"open_lag_days":0,"gap_closed_minus_open":238.46,"new_records":[{"d":"2026-10-13","s":1261.13,"m":"gpt-5.6-sol-xhigh","l":"openai"}],"claimed":0},{"id":"epoch-openbookqa","name":"OpenBookQA","source":"epoch","unit":"fraction","ceiling":1,"models":42,"as_of":"2024-04-23","best":{"d":"2024-04-23","s":0.88,"m":"phi-3-small 7.4B","l":"microsoft","weights":"open"},"saturation":0.88,"change_12m":0,"open_lag_days":null,"gap_closed_minus_open":-0.02,"new_records":[],"claimed":0},{"id":"epoch-deepresearch-bench","name":"DeepResearch Bench","source":"epoch","unit":"fraction","ceiling":1,"models":41,"as_of":"2026-05-28","best":{"d":"2026-02-05","s":0.553,"m":"Claude Opus 4.6 (high)","l":"anthropic","weights":"closed"},"saturation":0.553,"change_12m":0.027,"open_lag_days":93,"gap_closed_minus_open":0.202,"new_records":[],"claimed":0},{"id":"epoch-balrog","name":"Balrog","source":"epoch","unit":"fraction","ceiling":1,"models":40,"as_of":"2026-09-03","best":{"d":"2026-09-03","s":0.683,"m":"GPT-6 Astra (max)","l":"openai","weights":"closed"},"saturation":0.683,"change_12m":0.247,"open_lag_days":-64,"gap_closed_minus_open":0.334,"new_records":[],"claimed":0},{"id":"epoch-triviaqa","name":"TriviaQA","source":"epoch","unit":"fraction","ceiling":1,"models":38,"as_of":"2024-12-26","best":{"d":"2023-07-11","s":0.875,"m":"Claude 2","l":"anthropic","weights":"closed"},"saturation":0.875,"change_12m":0,"open_lag_days":618,"gap_closed_minus_open":0.046,"new_records":[],"claimed":0},{"id":"epoch-vpct","name":"VPCT","source":"epoch","unit":"fraction","ceiling":1,"models":38,"as_of":"2025-12-17","best":{"d":"2025-11-18","s":0.91,"m":"Gemini 3 Pro Preview","l":"google","weights":"closed"},"saturation":0.91,"change_12m":0.25,"open_lag_days":null,"gap_closed_minus_open":null,"new_records":[],"claimed":0},{"id":"epoch-blueprint-bench-2","name":"Blueprint-Bench 2","source":"epoch","unit":"fraction","ceiling":1,"models":35,"as_of":"2026-10-28","best":{"d":"2026-10-28","s":0.585683,"m":"Gemini 4 Carbon (unknown)","l":"google","weights":"closed"},"saturation":0.5857,"change_12m":0.585683,"open_lag_days":84,"gap_closed_minus_open":0.290672,"new_records":[{"d":"2026-09-22","s":0.511931,"m":"Claude Opus 5.5 (unknown)","l":"anthropic"},{"d":"2026-09-30","s":0.544469,"m":"Gemini 4 Argon (unknown)","l":"google"},{"d":"2026-10-28","s":0.585683,"m":"Gemini 4 Carbon (unknown)","l":"google"}],"claimed":0},{"id":"epoch-furniture-assembly","name":"Furniture Assembly","source":"epoch","unit":"fraction","ceiling":1,"models":35,"as_of":"2026-10-26","best":{"d":"2026-09-22","s":0.833333,"m":"Claude Opus 5.5 (max)","l":"anthropic","weights":"closed"},"saturation":0.8333,"change_12m":null,"open_lag_days":217,"gap_closed_minus_open":0.491666,"new_records":[{"d":"2026-09-22","s":0.833333,"m":"Claude Opus 5.5 (max)","l":"anthropic"}],"claimed":0},{"id":"epoch-swe-bench-verified","name":"SWE-Bench verified","source":"epoch","unit":"fraction","ceiling":1,"models":33,"as_of":"2026-06-16","best":{"d":"2026-04-16","s":0.834711,"m":"Claude Opus 4.7 (max)","l":"anthropic","weights":"closed"},"saturation":0.8347,"change_12m":0.099174,"open_lag_days":61,"gap_closed_minus_open":0.047711,"new_records":[],"claimed":0},{"id":"epoch-geobench","name":"GeoBench","source":"epoch","unit":"fraction","ceiling":1,"models":31,"as_of":"2025-12-17","best":{"d":"2025-12-17","s":0.88,"m":"Gemini 3 Flash Preview","l":"google","weights":"closed"},"saturation":0.88,"change_12m":0.07,"open_lag_days":63,"gap_closed_minus_open":0.26,"new_records":[],"claimed":0},{"id":"epoch-mirrorcode","name":"MirrorCode","source":"epoch","unit":"fraction","ceiling":1,"models":31,"as_of":"2026-10-12","best":{"d":"2026-09-28","s":0.977778,"m":"Claude Sonnet 5.5 (max)","l":"anthropic","weights":"closed"},"saturation":0.9778,"change_12m":0.933334,"open_lag_days":49,"gap_closed_minus_open":0.433334,"new_records":[{"d":"2026-09-28","s":0.977778,"m":"Claude Sonnet 5.5 (max)","l":"anthropic"}],"claimed":0},{"id":"epoch-surface-evolver-bench","name":"Surface Evolver Bench","source":"epoch","unit":"fraction","ceiling":1,"models":29,"as_of":"2026-10-06","best":{"d":"2026-06-09","s":0.95,"m":"Claude Fable 5 (high)","l":"anthropic","weights":"closed"},"saturation":0.95,"change_12m":0.7,"open_lag_days":37,"gap_closed_minus_open":0,"new_records":[],"claimed":0},{"id":"epoch-ebr-bench","name":"EBR-bench","source":"epoch","unit":"fraction","ceiling":1,"models":28,"as_of":"2026-10-26","best":{"d":"2026-10-26","s":0.780952,"m":"GPT-6.1 Terra (max)","l":"openai","weights":"closed"},"saturation":0.781,"change_12m":0.653968,"open_lag_days":313,"gap_closed_minus_open":0.685714,"new_records":[{"d":"2026-10-26","s":0.780952,"m":"GPT-6.1 Terra (max)","l":"openai"}],"claimed":0},{"id":"epoch-lambada","name":"LAMBADA","source":"epoch","unit":"fraction","ceiling":1,"models":26,"as_of":"2023-11-30","best":{"d":"2023-09-06","s":0.798,"m":"Falcon-180B","l":"technology-innovation-institute","weights":"open"},"saturation":0.798,"change_12m":0,"open_lag_days":null,"gap_closed_minus_open":-0.013,"new_records":[],"claimed":0},{"id":"epoch-weirdml-v3","name":"WeirdML v3","source":"epoch","unit":"fraction","ceiling":0.9,"models":25,"as_of":"2026-10-26","best":{"d":"2026-09-03","s":0.540453,"m":"GPT-6 Astra (xhigh)","l":"openai","weights":"closed"},"saturation":0.6005,"change_12m":null,"open_lag_days":130,"gap_closed_minus_open":0.341219,"new_records":[],"claimed":0},{"id":"swebench-full","name":"SWE-bench Full (agent submissions)","source":"swebench","unit":"fraction","ceiling":1,"models":24,"as_of":"2025-12-19","best":{"d":"2025-12-19","s":0.5262,"m":"Claude 4.5 Opus","l":"anthropic","weights":"closed"},"saturation":0.5262,"change_12m":0.0837,"open_lag_days":-2,"gap_closed_minus_open":0.3802,"new_records":[],"claimed":0},{"id":"epoch-gbaeval","name":"GBAEval","source":"epoch","unit":"fraction","ceiling":1,"models":23,"as_of":"2026-07-24","best":{"d":"2026-07-24","s":0.795985,"m":"Claude Opus 5 (unknown)","l":"anthropic","weights":"closed"},"saturation":0.796,"change_12m":null,"open_lag_days":149,"gap_closed_minus_open":0.312516,"new_records":[],"claimed":0},{"id":"epoch-scienceqa","name":"ScienceQA","source":"epoch","unit":"fraction","ceiling":1,"models":23,"as_of":"2024-08-16","best":{"d":"2024-08-16","s":0.913,"m":"Phi-3.5-vision-instruct","l":"unknown","weights":"unknown"},"saturation":0.913,"change_12m":0,"open_lag_days":-332,"gap_closed_minus_open":0.1433,"new_records":[],"claimed":0},{"id":"epoch-cl-bench","name":"CL-bench","source":"epoch","unit":"fraction","ceiling":1,"models":22,"as_of":"2026-03-31","best":{"d":"2026-03-05","s":0.279,"m":"GPT-5.4 (xhigh)","l":"openai","weights":"closed"},"saturation":0.279,"change_12m":0.101,"open_lag_days":75,"gap_closed_minus_open":0.086,"new_records":[],"claimed":0},{"id":"epoch-cybench","name":"Cybench","source":"epoch","unit":"fraction","ceiling":1,"models":22,"as_of":"2026-02-05","best":{"d":"2026-02-05","s":0.93,"m":"Claude Opus 4.6 (unknown thinking)","l":"anthropic","weights":"closed"},"saturation":0.93,"change_12m":0.33,"open_lag_days":62,"gap_closed_minus_open":0.855,"new_records":[],"claimed":0},{"id":"swebench-multimodal","name":"SWE-bench Multimodal (agent submissions)","source":"swebench","unit":"fraction","ceiling":1,"models":22,"as_of":"2025-11-17","best":{"d":"2025-07-01","s":0.3598,"m":"o3","l":"openai","weights":"closed"},"saturation":0.3598,"change_12m":0,"open_lag_days":null,"gap_closed_minus_open":null,"new_records":[],"claimed":0},{"id":"epoch-frontierswe","name":"FrontierSWE","source":"epoch","unit":"fraction","ceiling":1,"models":21,"as_of":"2026-10-07","best":{"d":"2026-09-03","s":0.655076,"m":"GPT-6 Astra (max)","l":"openai","weights":"closed"},"saturation":0.6551,"change_12m":null,"open_lag_days":66,"gap_closed_minus_open":0.353257,"new_records":[],"claimed":0},{"id":"epoch-algotune","name":"AlgoTune","source":"epoch","unit":"points","ceiling":null,"models":18,"as_of":"2026-03-05","best":{"d":"2025-12-11","s":2.05,"m":"GPT-5.2 (medium)","l":"openai","weights":"closed"},"saturation":null,"change_12m":0.33,"open_lag_days":-86,"gap_closed_minus_open":0.35,"new_records":[],"claimed":0},{"id":"epoch-remote-labor-index","name":"Remote Labor Index","source":"epoch","unit":"fraction","ceiling":1,"models":18,"as_of":"2026-09-30","best":{"d":"2026-09-22","s":0.2125,"m":"Claude Opus 5.5 (unknown)","l":"anthropic","weights":"closed"},"saturation":0.2125,"change_12m":0.1917,"open_lag_days":null,"gap_closed_minus_open":null,"new_records":[{"d":"2026-09-22","s":0.2125,"m":"Claude Opus 5.5 (unknown)","l":"anthropic"}],"claimed":0},{"id":"epoch-cl-bench-life","name":"CL-bench Life","source":"epoch","unit":"fraction","ceiling":1,"models":16,"as_of":"2026-04-24","best":{"d":"2026-04-23","s":0.222,"m":"GPT-5.5 (high)","l":"openai","weights":"closed"},"saturation":0.222,"change_12m":0.127,"open_lag_days":162,"gap_closed_minus_open":0.087,"new_records":[],"claimed":0},{"id":"epoch-cadeval","name":"CadEval","source":"epoch","unit":"fraction","ceiling":1,"models":14,"as_of":"2025-04-16","best":{"d":"2025-04-16","s":0.74,"m":"o3 (medium)","l":"openai","weights":"closed"},"saturation":0.74,"change_12m":0,"open_lag_days":null,"gap_closed_minus_open":null,"new_records":[],"claimed":0},{"id":"epoch-frontiermath-erdos","name":"FrontierMath-Erdos","source":"epoch","unit":"fraction","ceiling":1,"models":14,"as_of":"2026-09-30","best":{"d":"2026-09-29","s":0.044118,"m":"GPT-6.1 Sol (max)","l":"openai","weights":"closed"},"saturation":0.0441,"change_12m":null,"open_lag_days":84,"gap_closed_minus_open":0.044118,"new_records":[{"d":"2026-09-29","s":0.044118,"m":"GPT-6.1 Sol (max)","l":"openai"}],"claimed":0},{"id":"epoch-osworld-2-0","name":"OSWorld 2.0","source":"epoch","unit":"fraction","ceiling":1,"models":14,"as_of":"2026-07-24","best":{"d":"2026-07-24","s":0.3143,"m":"Claude Opus 5 (max)","l":"anthropic","weights":"closed"},"saturation":0.3143,"change_12m":null,"open_lag_days":62,"gap_closed_minus_open":0.2683,"new_records":[],"claimed":0},{"id":"epoch-the-agent-company","name":"The Agent Company","source":"epoch","unit":"fraction","ceiling":1,"models":14,"as_of":"2025-09-29","best":{"d":"2025-09-29","s":0.429,"m":"DeepSeek-V3.2-Exp","l":"deepseek","weights":"open"},"saturation":0.429,"change_12m":0,"open_lag_days":null,"gap_closed_minus_open":-0.018,"new_records":[],"claimed":0},{"id":"swebench-multilingual","name":"SWE-bench Multilingual (agent submissions)","source":"swebench","unit":"fraction","ceiling":1,"models":13,"as_of":"2026-02-20","best":{"d":"2026-02-13","s":0.727,"m":"Gemini 3 Flash","l":"google","weights":"closed"},"saturation":0.727,"change_12m":null,"open_lag_days":0,"gap_closed_minus_open":0.03,"new_records":[],"claimed":0},{"id":"epoch-posttrainbench","name":"PostTrainBench","source":"epoch","unit":"fraction","ceiling":1,"models":12,"as_of":"2026-07-24","best":{"d":"2026-06-09","s":0.4179,"m":"Claude Fable 5 (max)","l":"anthropic","weights":"closed"},"saturation":0.4179,"change_12m":null,"open_lag_days":49,"gap_closed_minus_open":0.0983,"new_records":[],"claimed":0},{"id":"epoch-anli","name":"ANLI","source":"epoch","unit":"fraction","ceiling":1,"models":11,"as_of":"2024-04-23","best":{"d":"2023-11-06","s":0.581,"m":"GPT-3.5 Turbo (Nov 2023)","l":"openai","weights":"closed"},"saturation":0.581,"change_12m":0,"open_lag_days":169,"gap_closed_minus_open":0,"new_records":[],"claimed":0},{"id":"epoch-gdpval","name":"GDPval","source":"epoch","unit":"fraction","ceiling":1,"models":11,"as_of":"2025-12-11","best":{"d":"2025-12-11","s":0.497,"m":"GPT-5.2 (none)","l":"openai","weights":"closed"},"saturation":0.497,"change_12m":0.061,"open_lag_days":null,"gap_closed_minus_open":null,"new_records":[],"claimed":0},{"id":"epoch-exploitbench","name":"ExploitBench","source":"epoch","unit":"fraction","ceiling":1,"models":9,"as_of":"2026-04-23","best":{"d":"2026-04-07","s":0.706,"m":"Claude Mythos Preview (Early)","l":"anthropic","weights":"closed"},"saturation":0.706,"change_12m":0.572,"open_lag_days":62,"gap_closed_minus_open":0.543,"new_records":[],"claimed":0},{"id":"epoch-osworld","name":"OSWorld","source":"epoch","unit":"fraction","ceiling":1,"models":9,"as_of":"2026-02-17","best":{"d":"2026-02-17","s":0.721,"m":"Claude Sonnet 4.6 (no thinking)","l":"anthropic","weights":"closed"},"saturation":0.721,"change_12m":0.292,"open_lag_days":64,"gap_closed_minus_open":0.088,"new_records":[],"claimed":0},{"id":"epoch-spatialviz-bench","name":"SpatialViz-Bench","source":"epoch","unit":"fraction","ceiling":1,"models":8,"as_of":"2025-06-17","best":{"d":"2025-06-17","s":0.4466,"m":"Gemini 2.5 Pro (Jun 2025)","l":"google","weights":"closed"},"saturation":0.4466,"change_12m":0,"open_lag_days":109,"gap_closed_minus_open":0.1042,"new_records":[],"claimed":0},{"id":"epoch-spireclimb","name":"SpireClimb","source":"epoch","unit":"fraction","ceiling":1,"models":7,"as_of":"2026-09-29","best":{"d":"2026-09-03","s":0.256667,"m":"GPT-6 Astra (max)","l":"openai","weights":"closed"},"saturation":0.2567,"change_12m":null,"open_lag_days":null,"gap_closed_minus_open":null,"new_records":[],"claimed":0},{"id":"epoch-spireclimb-slay-the-spire-1-ascension-20-heart","name":"SpireClimb (Slay the Spire 1, Ascension 20 Heart)","source":"epoch","unit":"fraction","ceiling":1,"models":7,"as_of":"2026-09-29","best":{"d":"2026-09-29","s":0.2,"m":"GPT-6.1 Sol (max)","l":"openai","weights":"closed"},"saturation":0.2,"change_12m":null,"open_lag_days":null,"gap_closed_minus_open":null,"new_records":[{"d":"2026-09-29","s":0.2,"m":"GPT-6.1 Sol (max)","l":"openai"}],"claimed":0},{"id":"epoch-spireclimb-slay-the-spire-2-ascension-10","name":"SpireClimb (Slay the Spire 2, Ascension 10)","source":"epoch","unit":"fraction","ceiling":1,"models":7,"as_of":"2026-09-29","best":{"d":"2026-09-03","s":0.346667,"m":"GPT-6 Astra (max)","l":"openai","weights":"closed"},"saturation":0.3467,"change_12m":null,"open_lag_days":null,"gap_closed_minus_open":null,"new_records":[],"claimed":0},{"id":"epoch-mindcube","name":"MindCube","source":"epoch","unit":"fraction","ceiling":1,"models":5,"as_of":"2025-05-22","best":{"d":"2025-03-12","s":0.4667,"m":"Gemma 3 12B","l":"google","weights":"open"},"saturation":0.4667,"change_12m":0,"open_lag_days":null,"gap_closed_minus_open":-0.0192,"new_records":[],"claimed":0},{"id":"epoch-csqa2","name":"CSQA2","source":"epoch","unit":"fraction","ceiling":1,"models":3,"as_of":"2023-07-18","best":{"d":"2023-06-13","s":0.57,"m":"GPT-3.5 Turbo (June 2023)","l":"openai","weights":"closed"},"saturation":0.57,"change_12m":0,"open_lag_days":null,"gap_closed_minus_open":null,"new_records":[],"claimed":0},{"id":"epoch-superglue","name":"SuperGLUE","source":"epoch","unit":"fraction","ceiling":1,"models":1,"as_of":"2022-01-27","best":{"d":"2022-01-27","s":0.718,"m":"InstructGPT 175B","l":"openai","weights":"closed"},"saturation":0.718,"change_12m":0,"open_lag_days":null,"gap_closed_minus_open":null,"new_records":[],"claimed":0},{"id":"claimed-sec-bench-pro","name":"SEC-bench Pro","source":"self_reported","unit":"fraction","ceiling":1,"models":0,"as_of":"","best":null,"saturation":null,"change_12m":null,"open_lag_days":null,"gap_closed_minus_open":null,"new_records":[],"claimed":1}]},"meta":{"as_of":"2026-10-31T00:59:47+00:00","source":"ohmega.ai-benchmarks/1","version":"v1","cite_as":"https://ohmega.monster/intel/ai","license":"Ohmega claims, summaries, grades, Pulses and forecasts may be quoted with a link to cite_as. Source texts belong to their publishers; follow the source URLs. No warranty: readings are our assessment as of as_of, not guarantees.","verification":"measured","note":"Scores are as published by each source (Epoch AI, CC BY 4.0; SWE-bench; LMArena) and are independently measured. Frontier = best measured score so far, placed by model release date (Elo ratings: by leaderboard publish date, captured since our first fetch). Fractions are 0..1. `claimed` rows are lab-claimed scores from blogs and model cards: unverified, attribute them to the lab. Check `sources[].stale` before relying on freshness; each benchmark's `as_of` is its newest result."}}