⌕Is ___ SOTA?
benchmarks / arc-agi

ARC-AGI

score. Status active. Contamination risk med. Maintainer ARC Prize / Epoch collection. Source.

ModelScoreDateKind
Claude Fable 50.9852026-06-09observation
Claude Fable 50.9852026-06-09observation
GPT-6 Astra0.9852026-09-03observation
GPT-6 Astra0.9852026-09-03observation
Claude Opus 5.50.9852026-09-22observation
GPT-6.1 Sol0.9852026-09-29observation
GPT-6.1 Sol0.9852026-09-29observation
Gemini 3.8 Flash0.9852026-09-02observation
Gemini 3.1 Pro Preview0.982026-02-19observation
GPT-5.6 Sol0.9752026-07-09observation
Claude Opus 50.9752026-07-24observation
Claude Opus 50.9752026-07-24observation
Claude Fable 5.10.9752026-09-01observation
GPT-6 Astra0.9752026-09-03observation
GPT-6 Astra0.9752026-09-03observation
Claude Opus 5.50.9752026-09-22observation
Claude Opus 5.50.9752026-09-22observation
Claude Opus 5.50.9752026-09-22observation
Gemini 3.8 Flash0.9752026-09-02observation
GPT-5.6 Sol0.972026-07-09observation
GPT-5.5 Pro0.9652026-04-23observation
GPT-5.6 Sol0.9652026-07-09observation
GPT-5.6 Terra0.9652026-07-09observation
Claude Fable 5.10.9652026-09-01observation
GPT-6 Astra0.9652026-09-03observation
GPT-6.1 Sol0.9652026-09-29observation
Gemini 3 Deep Think Preview0.962026-02-12observation
Claude Fable 5.10.962026-09-01observation
Claude Fable 50.95500000000000012026-06-09observation
Gemini 3.7 Flash0.95500000000000012026-08-13observation