Skip to content

OpenAI

GPT-6.1 Sol

Close to GPT-6 Astra's ability at one-fifth of its API price. Scores 100% on Epoch AI's FrontierMath Tier 4.

Announced
Access
Public

Benchmark scores

TestScoreBest on the trackerHow it was measured
Epoch Capabilities Indexmany tests combined into one score169 pointsGPT-6.2 Astra 174 pointsIndependentlisted as "GPT-6.1 Sol", CI 165.9 to 172.24Run 9 Nov 2026
AA Intelligence Index v5about ten tests in one score54.6 pointsClaude Fable 5.2 64.6 pointsIndependentv5.0.1; listed as GPT-6.1 Sol (Max)Run 10 Nov 2026
Terminal-Bench 4.1long command-line tasks61%Claude Opus 5.5 66%IndependentCodex harness, max effort, 330 trials, ±2.97 (95% CI)
APEX-Agentsprofessional office tasks60%Gemini 4 Carbon 84%Independentpass@1, max effort; listed as "GPT-6.1 Sol"
ARC-AGI-3puzzle games, standard harnessTypical first-time human player: 100%53%Claude Fable 5.2 81%IndependentARC Prize config "GPT-6.1 Sol (Max)" (openai-gpt-6-1-sol-max), standard harness
FrontierMath T1–3hard maths problems94%Claude Fable 5.2 95%Independentmax reasoning effort (gpt-6.1-sol_max)Run 29 Sept 2026
GPQA DiamondPhD-level science questionsPhD experts: 70%Saturated95%Gemini 4 Argon 96%Independentmax reasoning effort (gpt-6.1-sol_max)Run 29 Sept 2026
FrontierMath T4research-level maths problemsSaturated100%Top scoreIndependentmax reasoning effort (gpt-6.1-sol_max)Run 29 Sept 2026
ARC-AGI-2visual puzzles, model aloneAverage person: 60%Saturated94%GPT-6 Astra 95%IndependentARC Prize config "GPT-6.1 Sol (Max)" (openai-gpt-6-1-sol-max)
OTIS Mock AIMEcompetition mathsSaturated100%Top scoreIndependentmax reasoning effort (gpt-6.1-sol_max)Run 29 Sept 2026

Compare every model on the same tests →

Also on these charts

Sources