Acropolis

GPT-5.5 Pro

OpenAIGrade C. Evidence grade C of A to D: thin.ProprietaryReleased 23 Apr 2026
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

#42 / 299
84.6%$10.51StandardNot stated
ARC-AGI-2

Epoch AI Benchmarking Hub · xhigh

#44 / 299
84.2%$10.76StandardNot stated
ARC-AGI-2

Epoch AI Benchmarking Hub · xhigh

#44 / 299
84.2%UnknownStandardNot 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.
    high

    84.6

    $10.51

    xhigh

    84.2