Acropolis

GPT-6 Sol

OpenAIGrade A. Evidence grade A of A to D: strongest.ProprietaryReleased 22 Sept 2026Consistent across benchmarks
Compare with peers

Domain profile

Median within-board percentile by capability domain across boards with sufficient peer overlap.

  • Agentic (4 boards): 82 (scale 0 to 100)Agentic (4 boards)82
  • Coding (4 boards): 82 (scale 0 to 100)Coding (4 boards)82
  • Finance (1 boards): 97 (scale 0 to 100)Finance (1 boards)97
  • General text (3 boards): 85 (scale 0 to 100)General text (3 boards)85
  • Science (2 boards): 93 (scale 0 to 100)Science (2 boards)93
  • Search (2 boards): 65 (scale 0 to 100)Search (2 boards)65
  • Vision (2 boards): 66 (scale 0 to 100)Vision (2 boards)66

Benchmark appearances

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

BenchmarkRankScoreCostStatusObserved
Vending-Bench 2

Epoch AI Benchmarking Hub · Default

#3 / 72
$14,427.85UnknownStandardNot stated
LMArena Search (style control)

LMArena · Overall · max

#4 / 42
1,229[1,221 to 1,237]UnknownStandard9 Nov 2026
LMArena WebDev

LMArena · Image to webdev · max

#7 / 64
1,657[1,646 to 1,667]UnknownStandard15 Nov 2026
FrontierCode

Epoch AI Benchmarking Hub · max · Codex CLI

#9 / 51
49.3%UnknownStandardNot stated
GPQA diamond

Epoch AI Benchmarking Hub · max

#11 / 338Tied #1
94.3%[91.4% to 97.2%]UnknownStandard22 Sept 2026
LMArena WebDev

LMArena · Webdev HTML · max

#13 / 155
1,683[1,664 to 1,702]UnknownStandard15 Nov 2026
LMArena Agent

LMArena · Overall · max

#15 / 64
0.1[0 to 0.1]$1.85Standard15 Nov 2026
LMArena WebDev

LMArena · Overall · max

#16 / 156
1,673[1,665 to 1,681]UnknownStandard15 Nov 2026
LMArena WebDev

LMArena · Webdev · max

#16 / 156
1,673[1,665 to 1,681]UnknownStandard15 Nov 2026
LMArena WebDev

LMArena · Webdev react · max

#17 / 140
1,675[1,665 to 1,685]UnknownStandard15 Nov 2026
ARC-AGI-2

Epoch AI Benchmarking Hub · max

#23 / 299
89.6%UnknownStandardNot stated
LMArena Search

LMArena · Overall · max

#26 / 42
1,174[1,167 to 1,182]UnknownStandard9 Nov 2026
LMArena Vision (style control)

LMArena · Homework · max

#26 / 118
1,310[1,289 to 1,331]UnknownStandard9 Nov 2026
LMArena Text (style control)

LMArena · Coding · max

#34 / 416
1,521[1,511 to 1,530]UnknownStandard11 Nov 2026
LMArena Vision (style control)

LMArena · Chinese · max

#37 / 135
1,313[1,275 to 1,351]UnknownStandard9 Nov 2026
LMArena Vision (style control)

LMArena · Diagram · max

#37 / 123
1,298[1,280 to 1,315]UnknownStandard9 Nov 2026
LMArena Vision (style control)

LMArena · OCR · max

#42 / 123
1,278[1,267 to 1,289]UnknownStandard9 Nov 2026
LMArena Text (style control)

LMArena · Polish · max

#45 / 243
1,474[1,440 to 1,507]UnknownStandard11 Nov 2026
LMArena Vision

LMArena · Homework · max

#45 / 118
1,290[1,270 to 1,311]UnknownStandard9 Nov 2026
LMArena Text (style control)

LMArena · Industry software and IT services · max

#47 / 421
1,506[1,498 to 1,514]UnknownStandard11 Nov 2026
LMArena Vision (style control)

LMArena · Overall · max

#47 / 167
1,258[1,248 to 1,268]UnknownStandard9 Nov 2026
Bug Hunt Bench

Bug Hunt Bench · max · Codex CLI

#53 / 145
29.3$9.33Standard23 Sept 2026
LMArena Text (style control)

LMArena · Expert · max

#53 / 371
1,496[1,482 to 1,510]UnknownStandard11 Nov 2026
LMArena Vision (style control)

LMArena · English · max

#54 / 167
1,251[1,234 to 1,267]UnknownStandard9 Nov 2026
ARC-AGI-2

Epoch AI Benchmarking Hub · xhigh

#55 / 299
78.1%UnknownStandardNot stated
LMArena Text (style control)

LMArena · Russian · max

#59 / 385
1,465[1,453 to 1,477]UnknownStandard11 Nov 2026
LMArena Vision (style control)

LMArena · Creative writing vision · max

#59 / 106
1,232[1,200 to 1,263]UnknownStandard9 Nov 2026
LMArena Text (style control)

LMArena · Korean · max

#61 / 289
1,420[1,385 to 1,454]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Hard prompts · max

#62 / 421
1,483[1,477 to 1,490]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Exclude ties · max

#64 / 421
1,463[1,456 to 1,471]UnknownStandard11 Nov 2026
LMArena Vision

LMArena · Diagram · max

#64 / 123
1,264[1,247 to 1,281]UnknownStandard9 Nov 2026
LMArena Text (style control)

LMArena · Non English · max

#67 / 421
1,445[1,439 to 1,452]UnknownStandard11 Nov 2026
Bug Hunt Bench

Bug Hunt Bench · xhigh · Codex CLI

#68 / 145
25$7.67Standard23 Sept 2026
LMArena Text (style control)

LMArena · Multi turn · max

#68 / 419
1,468[1,457 to 1,480]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Overall · max

#68 / 421
1,457[1,452 to 1,463]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry legal and government · max

#69 / 393
1,465[1,449 to 1,480]UnknownStandard11 Nov 2026
LMArena Vision

LMArena · OCR · max

#69 / 123
1,251[1,240 to 1,262]UnknownStandard9 Nov 2026
LMArena Text (style control)

LMArena · Hard prompts English · max

#70 / 419
1,484[1,474 to 1,494]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Instruction following · max

#70 / 421
1,449[1,440 to 1,457]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Longer query · max

#70 / 399
1,466[1,458 to 1,473]UnknownStandard11 Nov 2026
LMArena Vision (style control)

LMArena · Humor · max

#70 / 102
1,209[1,165 to 1,253]UnknownStandard9 Nov 2026
LMArena Text (style control)

LMArena · French · max

#71 / 301
1,477[1,452 to 1,501]UnknownStandard11 Nov 2026
LMArena Vision

LMArena · Chinese · max

#71 / 135
1,271[1,233 to 1,309]UnknownStandard9 Nov 2026
LMArena Text (style control)

LMArena · German · max

#73 / 320
1,453[1,419 to 1,487]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Japanese · max

#73 / 288
1,421[1,387 to 1,456]UnknownStandard11 Nov 2026
ARC-AGI-2

Epoch AI Benchmarking Hub · high

#76 / 299
68.9%UnknownStandardNot stated
LMArena Text (style control)

LMArena · Creative writing · max

#77 / 419
1,428[1,418 to 1,438]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry writing and literature and language · max

#77 / 420
1,435[1,426 to 1,444]UnknownStandard11 Nov 2026
LMArena Vision

LMArena · Overall · max

#77 / 167
1,233[1,223 to 1,242]UnknownStandard9 Nov 2026
LMArena Text (style control)

LMArena · Industry entertainment and sports and media · max

#78 / 419
1,425[1,416 to 1,435]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Spanish · max

#78 / 304
1,449[1,426 to 1,471]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry business and management and financial operations · max

#81 / 414
1,450[1,439 to 1,460]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry mathematical · max

#81 / 402
1,455[1,438 to 1,473]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · English · max

#84 / 421
1,460[1,452 to 1,468]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry life and physical and social science · max

#85 / 419
1,468[1,457 to 1,480]UnknownStandard11 Nov 2026
LMArena Vision

LMArena · Creative writing vision · max

#85 / 106
1,197[1,167 to 1,228]UnknownStandard9 Nov 2026
LMArena Vision

LMArena · English · max

#86 / 167
1,226[1,210 to 1,242]UnknownStandard9 Nov 2026
Bug Hunt Bench

Bug Hunt Bench · high · Codex CLI

#88 / 145
20$4.16Standard23 Sept 2026
LMArena Vision

LMArena · Humor · max

#90 / 102
1,179[1,135 to 1,223]UnknownStandard9 Nov 2026
LMArena Text (style control)

LMArena · Industry medicine and healthcare · max

#97 / 391
1,459[1,442 to 1,476]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Math · max

#97 / 404
1,440[1,421 to 1,459]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Chinese · max

#103 / 395
1,477[1,461 to 1,493]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Coding · max

#106 / 416
1,449[1,440 to 1,459]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Expert · max

#107 / 371
1,439[1,425 to 1,452]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Instruction following · max

#107 / 421
1,410[1,402 to 1,418]UnknownStandard11 Nov 2026
ARC-AGI-2

Epoch AI Benchmarking Hub · medium

#112 / 299
57.8%UnknownStandardNot stated
Bug Hunt Bench

Bug Hunt Bench · medium · Codex CLI

#115 / 145
14$2.39Standard23 Sept 2026
LMArena Text

LMArena · Longer query · max

#122 / 399
1,415[1,407 to 1,422]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Russian · max

#122 / 385
1,408[1,396 to 1,420]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Multi turn · max

#131 / 419
1,414[1,403 to 1,426]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Polish · max

#132 / 243
1,397[1,363 to 1,431]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Japanese · max

#134 / 288
1,348[1,314 to 1,382]UnknownStandard11 Nov 2026
Bug Hunt Bench

Bug Hunt Bench · low · Codex CLI

#136 / 145
6$1.08Standard23 Sept 2026
LMArena Text

LMArena · Hard prompts · max

#138 / 421
1,419[1,413 to 1,426]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry software and IT services · max

#138 / 421
1,434[1,426 to 1,441]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry writing and literature and language · max

#139 / 420
1,387[1,378 to 1,396]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Creative writing · max

#145 / 419
1,370[1,360 to 1,381]UnknownStandard11 Nov 2026
LMArena Text

LMArena · German · max

#147 / 320
1,382[1,348 to 1,416]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry entertainment and sports and media · max

#147 / 419
1,367[1,357 to 1,376]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry mathematical · max

#147 / 402
1,414[1,396 to 1,432]UnknownStandard11 Nov 2026
ARC-AGI-2

Epoch AI Benchmarking Hub · low

#148 / 299
31.5%UnknownStandardNot stated
LMArena Text

LMArena · Korean · max

#148 / 289
1,343[1,309 to 1,377]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Hard prompts English · max

#152 / 419
1,418[1,408 to 1,427]UnknownStandard11 Nov 2026
LMArena Text

LMArena · French · max

#153 / 301
1,403[1,379 to 1,428]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Non English · max

#154 / 421
1,387[1,380 to 1,393]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry legal and government · max

#160 / 393
1,400[1,384 to 1,415]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Spanish · max

#160 / 304
1,390[1,367 to 1,412]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Exclude ties · max

#162 / 421
1,377[1,370 to 1,384]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Overall · max

#163 / 421
1,397[1,391 to 1,402]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Math · max

#168 / 404
1,396[1,377 to 1,415]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry business and management and financial operations · max

#170 / 414
1,383[1,372 to 1,393]UnknownStandard11 Nov 2026
LMArena Text

LMArena · English · max

#171 / 421
1,397[1,389 to 1,405]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Chinese · max

#178 / 395
1,405[1,390 to 1,421]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry life and physical and social science · max

#181 / 419
1,386[1,375 to 1,397]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry medicine and healthcare · max

#197 / 391
1,366[1,349 to 1,383]UnknownStandard11 Nov 2026
ARC-AGI-2

Epoch AI Benchmarking Hub · none

#250 / 299
1.7%UnknownStandardNot stated

Reasoning effort scaling

Same benchmark and harness measured across multiple reasoning effort settings.

  • ARC-AGI-2

    Epoch AI Benchmarking Hub

    Monotone. Evidence grade A of A to D: strongest.
    none

    1.7

    low

    31.5

    medium

    57.8

    high

    68.9

    xhigh

    78.1

    max

    89.6

  • Bug Hunt Bench

    Bug Hunt Bench · Codex CLI

    Monotone. Evidence grade A of A to D: strongest.
    low

    6.0

    $1.08

    medium

    14.0

    $2.39

    high

    20.0

    $4.16

    xhigh

    25.0

    $7.67

    max

    29.3

    $9.33

Closest head-to-head rivals

Direct pairwise win rates for GPT-6 Sol across shared benchmark boards.