AI Capabilities and Benchmarking Hub

Filter

Show benchmarks (4)

Epoch AI–run benchmarks

FrontierMath Tier 4 (v2)FrontierMath Tiers 1-3 (v2)SWE-bench VerifiedMirrorCodeChess PuzzlesEarthborne Rangers (EBR-bench)Mystery Game PuzzlesSimpleQA VerifiedGPQA DiamondOTIS Mock AIME 2024-2025MATH Level 5

Benchmark creator–run benchmarks

Humanity's Last ExamARC-AGI-2ARC-AGI-1GDPvalRemote Labor IndexOSWorld 2.0OS WorldTerminal-Bench 2.0Aider PolyglotSimpleBenchVending-Bench 2WeirdML (v2)ForecastBenchSciCodeVideo-MMECritPtFrontierSWEGSOThe Agent CompanyAlgoTuneALE-BenchFiction.liveBenchEnigmaEvalBALROGVPCTAPEX-AgentsGDP.pdfExploitBenchPostTrainBenchFrontierCodeDeepSWECursorBenchGBAEvalBTF-3ProofBenchCL-benchCL-bench LifeBlueprint-Bench 2MindCubeSpatialViz-BenchCadEvalSurface Evolver Bench

Model developer–run benchmarks

LiveBenchCybenchAdversarial NLIARC AI2BBHBoolQCommonSenseQA 2GSM8KHellaSwagLAMBADAMMLUOpenBookQAPIQAScienceQASuperGLUETriviaQAWinoGrande

Benchmarking updates

Aug. 3, 2026

We've updated the MirrorCode leaderboard. Claude Fable 5 leads with a score of 64%, followed by GPT-5.6 Sol at 20%.

See the thread

Jul. 31, 2026

We've launched an expansion of FrontierMath: Open Problems! The benchmark now contains 50 significant, unsolved problems from research mathematics. AI has solved three so far.

See the thread

Jul. 28, 2026

AI has found a presentation for the absolute Galois group of the field of 2-adic numbers — the second problem solved in FrontierMath: Open Problems.

See the thread