Skip to content

OpenAI

GPT-4o (Nov 2024)

Announced
Access
Public

Benchmark scores

TestScoreBest on the trackerHow it was measured
Humanity's Last Examexpert questions, no tools2.7%GPT-6 Astra 55%Independentleaderboard entry "GPT-4o (November 2024)" (gpt-4o-2024-11-20)
GPQA DiamondPhD-level science questionsPhD experts: 70%Saturated48%Gemini 4 Argon 96%Independentdefault settings (gpt-4o-2024-11-20)Run 5 Feb 2025
ARC-AGI-2visual puzzles, model aloneAverage person: 60%Saturated0%GPT-6 Astra 95%IndependentARC Prize config "GPT-4o" (gpt-4o-2024-11-20)
OTIS Mock AIMEcompetition mathsSaturated6.3%GPT-5.5 100%Independentdefault settings (gpt-4o-2024-11-20)Run 25 Feb 2025
SWE-bench Verifiedfixing real software bugsKnown flaws31%Claude Opus 4.7 83%Independentdefault settings (gpt-4o-2024-11-20)Run 11 Feb 2026

Compare every model on the same tests →

Sources