Acropolis

GPT-5.1 Codex

OpenAIGrade B. Evidence grade B of A to D: good.ProprietaryReleased 13 Nov 2025
Compare with peers

Benchmark appearances

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

BenchmarkRankScoreCostStatusObserved
Terminal Bench

Epoch AI Benchmarking Hub · Codex CLI

#55 / 202
57.8%[52.1% to 63.5%]UnknownStandard16 Nov 2025
Terminal Bench

Epoch AI Benchmarking Hub · Crux

#56 / 202
57.8%[52% to 63.5%]UnknownStandard12 Nov 2025
Terminal Bench

Epoch AI Benchmarking Hub · Letta Code

#64 / 202
53.5%[47.9% to 59%]UnknownStandard12 Nov 2025
Terminal Bench

Epoch AI Benchmarking Hub · Terminus 2

#109 / 202
36.9%[30.6% to 43.2%]UnknownStandard17 Nov 2025
Terminal Bench

Epoch AI Benchmarking Hub · Terminus 2

#111 / 202
36.9%[30.5% to 43.2%]UnknownStandard12 Nov 2025
LMArena WebDev

LMArena · Webdev HTML · Default

#129 / 155
1,347[1,333 to 1,360]UnknownStandard15 Nov 2026
LMArena WebDev

LMArena · Overall · Default

#130 / 156
1,338[1,327 to 1,350]UnknownStandard15 Nov 2026
LMArena WebDev

LMArena · Webdev · Default

#130 / 156
1,338[1,327 to 1,350]UnknownStandard15 Nov 2026