Acropolis

o3-pro

OpenAIGrade C. Evidence grade C of A to D: thin.ProprietaryReleased 10 Jun 2025
Compare with peers

Benchmark appearances

Best verified run per board, with confidence intervals and Pareto frontier status.

BenchmarkRankScoreCostStatusObserved
ARC-AGI-2

Epoch AI Benchmarking Hub · high

#206 / 299
4.9%$7.55StandardNot stated
ARC-AGI-2

Epoch AI Benchmarking Hub · low

#238 / 299
2.1%$2.23StandardNot stated
ARC-AGI-2

Epoch AI Benchmarking Hub · medium

#241 / 299
1.9%$4.74StandardNot stated

Reasoning effort scaling

Same benchmark and harness measured across multiple reasoning effort settings.

  • ARC-AGI-2

    Epoch AI Benchmarking Hub

    1 regression. Evidence grade D of A to D: weakest.
    low

    2.0

    $2.23

    medium

    1.9

    $4.74

    high

    4.9

    $7.55