Skip to content

OpenAI

GPT-5.5

Aimed at complex professional work, released with GPT-5.5 Pro. In the API from 24 April 2026.

Announced
Released
Access
Public

Benchmark scores

TestScoreBest on the trackerHow it was measured
Epoch Capabilities Indexmany tests combined into one score160 pointsGPT-6.2 Astra 174 pointsIndependentlisted as "GPT-5.5", CI 158.02 to 162.38Run 9 Nov 2026
Remote Labor Indexpaid freelance projects6.3%Claude Fable 5.2 41%IndependentCodex GPT 5.5Run 1 Jul 2026
MirrorCoderebuilding whole programs10%Claude Opus 5.5 79%Independenthigh reasoning effort, MirrorCode (ML, +Private, 2L): 15 programs x 2 languages, 3 runs each, 10B-token / 7-day budget, ±6.0 SERun 10 Aug 2026
APEX-Agentsprofessional office tasks55%Gemini 4 Carbon 84%Independentpass@1, unknown effort; listed as "GPT-5.5"
ARC-AGI-3puzzle games, standard harnessTypical first-time human player: 100%0.4%Claude Fable 5.2 81%IndependentARC Prize config "GPT-5.5 (High)" (openai-gpt-5-5-2026-04-23-high), standard harness
FrontierMath T1–3hard maths problems85%Claude Fable 5.2 95%Independentxhigh reasoning effort (gpt-5.5_xhigh)Run 11 Jun 2026
GPQA DiamondPhD-level science questionsPhD experts: 70%Saturated94%Gemini 4 Argon 96%Independentxhigh reasoning effort, pre-release checkpoint (gpt-5.5-pre-release_xhigh)Run 24 Apr 2026
FrontierMath T4research-level maths problemsSaturated73%GPT-6.1 Sol 100%Independentxhigh reasoning effort (gpt-5.5_xhigh)Run 11 Jun 2026
ARC-AGI-2visual puzzles, model aloneAverage person: 60%Saturated85%GPT-6 Astra 95%IndependentARC Prize config "GPT-5.5 (XHigh)" (gpt-5-5-2026-04-22-thinking-xhigh)
OTIS Mock AIMEcompetition mathsSaturated100%Top scoreIndependentxhigh reasoning effort, pre-release checkpoint (gpt-5.5-pre-release_xhigh)Run 24 Apr 2026
SWE-bench Verifiedfixing real software bugsKnown flaws81%Claude Opus 4.7 83%Independentxhigh reasoning effort, pre-release checkpoint (gpt-5.5-pre-release_xhigh)Run 24 Apr 2026
METR time horizon (TH1.1 + TH2)task length an AI agent finishes half the time11 hGPT-6 Astra 1.7 daysIndependentPooled TH1.1 + TH2 suites; 95% CI 365.9–1116.4 min

Compare every model on the same tests →

Also on these charts

Sources