Acropolis

Opus 5

AnthropicGrade A. Evidence grade A of A to D: strongest.ProprietaryReleased 24 Jul 2026Consistent across benchmarks
Compare with peers

Domain profile

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

  • Agentic (5 boards): 77 (scale 0 to 100)Agentic (5 boards)77
  • Coding (5 boards): 77 (scale 0 to 100)Coding (5 boards)77
  • Finance (1 boards): 92 (scale 0 to 100)Finance (1 boards)92
  • General text (5 boards): 97 (scale 0 to 100)General text (5 boards)97
  • Science (2 boards): 92 (scale 0 to 100)Science (2 boards)92
  • Search (2 boards): 84 (scale 0 to 100)Search (2 boards)84
  • Vision (2 boards): 94 (scale 0 to 100)Vision (2 boards)94

Benchmark appearances

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

BenchmarkRankScoreCostStatusObserved
LMArena Text

LMArena · Spanish · max

#1 / 304Tied #1
1,522[1,503 to 1,541]UnknownStandard11 Nov 2026
LMArena Document

LMArena · Overall · high

#2 / 46Tied #1
1,509[1,503 to 1,514]UnknownStandard9 Nov 2026
LMArena Text

LMArena · French · high

#2 / 301
1,520[1,506 to 1,534]UnknownStandard11 Nov 2026
LMArena Text

LMArena · German · max

#2 / 320
1,520[1,495 to 1,545]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry mathematical · high

#2 / 402Tied #1
1,543[1,533 to 1,553]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Math · high

#2 / 404Tied #1
1,532[1,521 to 1,543]UnknownStandard11 Nov 2026
LMArena Vision

LMArena · Chinese · high

#2 / 135Tied #1
1,402[1,383 to 1,421]UnknownStandard9 Nov 2026
FrontierCode

Epoch AI Benchmarking Hub · medium · Claude Code

#3 / 51
53.4%UnknownStandardNot stated
LMArena Text

LMArena · Industry business and management and financial operations · max

#3 / 414Tied #1
1,499[1,491 to 1,507]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry medicine and healthcare · max

#3 / 391Tied #1
1,527[1,514 to 1,539]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Korean · max

#3 / 289Tied #1
1,504[1,481 to 1,528]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Math · max

#3 / 404Tied #1
1,531[1,516 to 1,545]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry mathematical · high

#3 / 402Tied #1
1,531[1,520 to 1,541]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Math · high

#3 / 404Tied #1
1,521[1,510 to 1,532]UnknownStandard11 Nov 2026
LMArena Vision (style control)

LMArena · Chinese · high

#3 / 135Tied #1
1,364[1,345 to 1,384]UnknownStandard9 Nov 2026
LMArena Text

LMArena · Industry business and management and financial operations · high

#4 / 414Tied #1
1,497[1,491 to 1,503]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry legal and government · max

#4 / 393Tied #1
1,528[1,516 to 1,539]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry life and physical and social science · max

#4 / 419Tied #1
1,533[1,524 to 1,541]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry mathematical · max

#4 / 402Tied #1
1,539[1,525 to 1,553]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Japanese · max

#4 / 288Tied #1
1,521[1,495 to 1,547]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Korean · high

#4 / 289Tied #1
1,502[1,484 to 1,520]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Polish · max

#4 / 243Tied #1
1,522[1,497 to 1,546]UnknownStandard11 Nov 2026
LMArena Search

LMArena · Overall · high

#5 / 42
1,239[1,232 to 1,246]UnknownStandard9 Nov 2026
LMArena Text

LMArena · Chinese · max

#5 / 395
1,575[1,562 to 1,587]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Exclude ties · max

#5 / 421
1,524[1,518 to 1,530]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Expert · max

#5 / 371Tied #1
1,552[1,542 to 1,563]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Hard prompts · max

#5 / 421
1,525[1,520 to 1,531]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry legal and government · high

#5 / 393Tied #1
1,521[1,513 to 1,530]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Non English · max

#5 / 421
1,500[1,495 to 1,506]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Overall · max

#5 / 421
1,507[1,502 to 1,511]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Polish · high

#5 / 243Tied #1
1,518[1,499 to 1,536]UnknownStandard11 Nov 2026
LMArena Vision

LMArena · Diagram · high

#5 / 123Tied #1
1,345[1,336 to 1,355]UnknownStandard9 Nov 2026
LMArena Vision

LMArena · Overall · high

#5 / 167Tied #1
1,317[1,311 to 1,324]UnknownStandard9 Nov 2026
LMArena Text

LMArena · Coding · max

#6 / 416
1,532[1,525 to 1,539]UnknownStandard11 Nov 2026
LMArena Text

LMArena · English · max

#6 / 421
1,507[1,501 to 1,513]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Expert · high

#6 / 371Tied #1
1,552[1,544 to 1,560]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Hard prompts English · max

#6 / 419
1,522[1,515 to 1,530]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry software and IT services · max

#6 / 421
1,527[1,520 to 1,533]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Non English · high

#6 / 421
1,495[1,491 to 1,499]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Spanish · high

#6 / 304Tied #1
1,505[1,492 to 1,519]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Korean · high

#6 / 289Tied #1
1,482[1,463 to 1,500]UnknownStandard11 Nov 2026
LMArena WebDev

LMArena · Image to webdev · max

#6 / 64
1,664[1,654 to 1,674]UnknownStandard15 Nov 2026
Vending-Bench 2

Epoch AI Benchmarking Hub · Default

#6 / 72
$11,181.87UnknownStandardNot stated
LMArena Text

LMArena · Chinese · high

#7 / 395
1,570[1,561 to 1,579]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Coding · high

#7 / 416
1,531[1,525 to 1,537]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Exclude ties · high

#7 / 421
1,518[1,513 to 1,523]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry life and physical and social science · high

#7 / 419
1,521[1,514 to 1,527]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry medicine and healthcare · high

#7 / 391
1,509[1,500 to 1,518]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry software and IT services · high

#7 / 421
1,526[1,521 to 1,531]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Instruction following · max

#7 / 421
1,518[1,511 to 1,524]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Longer query · max

#7 / 399
1,516[1,510 to 1,522]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Overall · high

#7 / 421
1,502[1,499 to 1,506]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Chinese · high

#7 / 395
1,549[1,540 to 1,558]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Expert · high

#7 / 371
1,536[1,528 to 1,544]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry mathematical · max

#7 / 402Tied #1
1,522[1,508 to 1,536]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Korean · max

#7 / 289Tied #1
1,481[1,458 to 1,505]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Math · max

#7 / 404Tied #1
1,515[1,500 to 1,529]UnknownStandard11 Nov 2026
LMArena Document (style control)

LMArena · Overall · high

#8 / 46
1,486[1,480 to 1,492]UnknownStandard9 Nov 2026
LMArena Text

LMArena · Creative writing · max

#8 / 419
1,494[1,486 to 1,502]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Hard prompts · high

#8 / 421
1,522[1,518 to 1,527]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Hard prompts English · high

#8 / 419
1,519[1,513 to 1,524]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry writing and literature and language · max

#8 / 420
1,499[1,492 to 1,506]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Instruction following · high

#8 / 421
1,515[1,510 to 1,521]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Japanese · high

#8 / 288
1,514[1,494 to 1,534]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Russian · max

#8 / 385
1,507[1,497 to 1,517]UnknownStandard11 Nov 2026
LMArena Vision

LMArena · OCR · high

#8 / 123
1,325[1,319 to 1,332]UnknownStandard9 Nov 2026
LMArena Text

LMArena · German · high

#9 / 320
1,502[1,483 to 1,521]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Longer query · high

#9 / 399
1,512[1,507 to 1,517]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Russian · high

#9 / 385
1,506[1,498 to 1,513]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Chinese · max

#9 / 395
1,547[1,534 to 1,559]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry medicine and healthcare · max

#9 / 391Tied #1
1,514[1,502 to 1,526]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Japanese · max

#9 / 288
1,508[1,482 to 1,534]UnknownStandard11 Nov 2026
LMArena Vision

LMArena · English · high

#9 / 167
1,310[1,301 to 1,319]UnknownStandard9 Nov 2026
LMArena Search (style control)

LMArena · Overall · high

#10 / 42
1,213[1,206 to 1,220]UnknownStandard9 Nov 2026
LMArena Text

LMArena · English · high

#10 / 421
1,503[1,497 to 1,508]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry entertainment and sports and media · max

#10 / 419
1,480[1,473 to 1,488]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Spanish · max

#10 / 304Tied #1
1,490[1,471 to 1,509]UnknownStandard11 Nov 2026
LMArena Vision

LMArena · Homework · high

#10 / 118
1,335[1,322 to 1,349]UnknownStandard9 Nov 2026
LMArena Vision (style control)

LMArena · Diagram · high

#10 / 123
1,326[1,316 to 1,335]UnknownStandard9 Nov 2026
LMArena Text

LMArena · Industry writing and literature and language · high

#11 / 420
1,494[1,488 to 1,500]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Multi turn · max

#11 / 419
1,499[1,490 to 1,508]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · German · max

#11 / 320
1,501[1,476 to 1,526]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Instruction following · high

#11 / 421
1,496[1,491 to 1,502]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Japanese · high

#11 / 288
1,506[1,486 to 1,526]UnknownStandard11 Nov 2026
LMArena WebDev

LMArena · Overall · max

#11 / 156
1,691[1,685 to 1,697]UnknownStandard15 Nov 2026
LMArena WebDev

LMArena · Webdev · max

#11 / 156
1,691[1,685 to 1,697]UnknownStandard15 Nov 2026
LMArena WebDev

LMArena · Webdev react · max

#11 / 140
1,713[1,706 to 1,720]UnknownStandard15 Nov 2026
LMArena Text

LMArena · Creative writing · high

#12 / 419
1,486[1,480 to 1,492]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry entertainment and sports and media · high

#12 / 419
1,477[1,471 to 1,483]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Instruction following · max

#12 / 421
1,494[1,488 to 1,501]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Multi turn · high

#13 / 419
1,494[1,487 to 1,500]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Expert · max

#13 / 371
1,530[1,520 to 1,541]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry life and physical and social science · max

#13 / 419
1,515[1,507 to 1,524]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Polish · high

#13 / 243Tied #1
1,502[1,483 to 1,520]UnknownStandard11 Nov 2026
LMArena Text

LMArena · French · max

#14 / 301
1,498[1,480 to 1,517]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · French · high

#14 / 301
1,507[1,493 to 1,522]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Hard prompts · high

#14 / 421
1,515[1,511 to 1,520]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry legal and government · max

#14 / 393
1,503[1,491 to 1,514]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Longer query · high

#14 / 399
1,502[1,497 to 1,507]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · English · high

#15 / 421
1,494[1,489 to 1,499]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Longer query · max

#15 / 399
1,502[1,495 to 1,508]UnknownStandard11 Nov 2026
LMArena Vision

LMArena · Creative writing vision · high

#15 / 106
1,309[1,294 to 1,324]UnknownStandard9 Nov 2026
LMArena WebDev

LMArena · Webdev HTML · max

#15 / 155
1,679[1,665 to 1,693]UnknownStandard15 Nov 2026
LMArena Text (style control)

LMArena · Creative writing · max

#16 / 419
1,473[1,465 to 1,481]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Exclude ties · high

#16 / 421
1,504[1,499 to 1,509]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry writing and literature and language · max

#16 / 420
1,481[1,474 to 1,488]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Overall · high

#16 / 421
1,490[1,486 to 1,493]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Polish · max

#16 / 243Tied #1
1,501[1,476 to 1,525]UnknownStandard11 Nov 2026
LMArena Vision (style control)

LMArena · Homework · high

#16 / 118
1,321[1,308 to 1,335]UnknownStandard9 Nov 2026
LMArena Vision (style control)

LMArena · OCR · high

#16 / 123
1,300[1,293 to 1,307]UnknownStandard9 Nov 2026
LMArena WebDev

LMArena · Webdev react · high

#16 / 140
1,678[1,672 to 1,684]UnknownStandard15 Nov 2026
ARC-AGI-2

Epoch AI Benchmarking Hub · max

#17 / 299
90.4%UnknownStandardNot stated
LMArena Agent

LMArena · Overall · max

#17 / 64
0.1[0 to 0.1]$4.50Standard15 Nov 2026
LMArena Text (style control)

LMArena · Hard prompts · max

#17 / 421
1,513[1,508 to 1,519]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry legal and government · high

#17 / 393
1,500[1,491 to 1,509]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Non English · max

#17 / 421
1,480[1,474 to 1,485]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Russian · high

#17 / 385
1,496[1,489 to 1,504]UnknownStandard11 Nov 2026
LMArena Vision

LMArena · Humor · high

#17 / 102
1,294[1,274 to 1,313]UnknownStandard9 Nov 2026
LMArena Text (style control)

LMArena · Exclude ties · max

#18 / 421
1,503[1,497 to 1,509]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Hard prompts English · high

#18 / 419
1,514[1,508 to 1,519]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry life and physical and social science · high

#18 / 419
1,510[1,503 to 1,516]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry writing and literature and language · high

#18 / 420
1,481[1,475 to 1,487]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Non English · high

#18 / 421
1,479[1,475 to 1,484]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · English · max

#19 / 421
1,493[1,487 to 1,499]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Hard prompts English · max

#19 / 419
1,512[1,504 to 1,519]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Overall · max

#19 / 421
1,489[1,484 to 1,494]UnknownStandard11 Nov 2026
LMArena Vision (style control)

LMArena · Overall · high

#19 / 167
1,285[1,279 to 1,292]UnknownStandard9 Nov 2026
GPQA diamond

Epoch AI Benchmarking Hub · max

#20 / 338Tied #1
93.9%[91% to 96.8%]UnknownStandard24 Jul 2026
LMArena Text (style control)

LMArena · Coding · high

#20 / 416
1,532[1,526 to 1,538]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Creative writing · high

#20 / 419
1,471[1,464 to 1,477]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry entertainment and sports and media · high

#20 / 419
1,464[1,458 to 1,470]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry medicine and healthcare · high

#20 / 391
1,504[1,495 to 1,513]UnknownStandard11 Nov 2026
LMArena WebDev

LMArena · Overall · high

#20 / 156
1,657[1,651 to 1,662]UnknownStandard15 Nov 2026
LMArena WebDev

LMArena · Webdev · high

#20 / 156
1,657[1,651 to 1,662]UnknownStandard15 Nov 2026
LMArena Text (style control)

LMArena · Industry business and management and financial operations · high

#21 / 414
1,485[1,479 to 1,492]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry software and IT services · high

#21 / 421
1,522[1,517 to 1,527]UnknownStandard11 Nov 2026
LMArena WebDev

LMArena · Webdev HTML · high

#21 / 155
1,650[1,639 to 1,662]UnknownStandard15 Nov 2026
LMArena Text (style control)

LMArena · Industry entertainment and sports and media · max

#22 / 419
1,462[1,455 to 1,469]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Russian · max

#22 / 385
1,492[1,483 to 1,502]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Spanish · high

#22 / 304
1,478[1,464 to 1,491]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry software and IT services · max

#24 / 421
1,517[1,511 to 1,524]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · German · high

#26 / 320
1,488[1,469 to 1,507]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Coding · max

#27 / 416
1,528[1,521 to 1,536]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry business and management and financial operations · max

#27 / 414
1,483[1,475 to 1,491]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Multi turn · max

#27 / 419
1,486[1,477 to 1,495]UnknownStandard11 Nov 2026
LMArena Vision (style control)

LMArena · Creative writing vision · high

#27 / 106
1,271[1,256 to 1,286]UnknownStandard9 Nov 2026
LMArena Vision (style control)

LMArena · English · high

#27 / 167
1,277[1,268 to 1,286]UnknownStandard9 Nov 2026
LMArena Text (style control)

LMArena · Multi turn · high

#28 / 419
1,486[1,479 to 1,493]UnknownStandard11 Nov 2026
ARC-AGI-2

Epoch AI Benchmarking Hub · max

#29 / 299
88.3%UnknownStandardNot stated
ARC-AGI-2

Epoch AI Benchmarking Hub · high

#29 / 299
88.3%UnknownStandardNot stated
LMArena Vision (style control)

LMArena · Humor · high

#29 / 102
1,265[1,246 to 1,285]UnknownStandard9 Nov 2026
CursorBench

CursorBench · max · Cursor

#30 / 91
46.6%$11.95StandardNot stated
GPQA diamond

Epoch AI Benchmarking Hub · Default

#31 / 338Tied #1
92.9%[89.3% to 96.5%]UnknownStandard6 Aug 2026
CursorBench

CursorBench · xhigh · Cursor

#33 / 91
46.1%$11.43StandardNot stated
CursorBench

CursorBench · high · Cursor

#39 / 91
44.7%$9.00StandardNot stated
CursorBench

CursorBench · medium · Cursor

#43 / 91
43.3%$6.94StandardNot stated
CursorBench

CursorBench · low · Cursor

#52 / 91
40.7%$4.87StandardNot stated
LMArena Text (style control)

LMArena · French · max

#55 / 301
1,481[1,462 to 1,500]UnknownStandard11 Nov 2026
Bug Hunt Bench

Bug Hunt Bench · max · Claude Code

#62 / 145
27$54.94Standard1 Aug 2026
Bug Hunt Bench

Bug Hunt Bench · xhigh · Claude Code

#65 / 145
26$63.25Standard28 Aug 2026
Bug Hunt Bench

Bug Hunt Bench · medium · Claude Code

#70 / 145
24$37.40Standard27 Aug 2026
GPQA diamond

Epoch AI Benchmarking Hub · low

#82 / 338
87.9%[83.3% to 92.4%]UnknownStandard6 Aug 2026
Bug Hunt Bench

Bug Hunt Bench · high · Claude Code

#84 / 145
21$41.51Standard26 Jul 2026

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.
    high

    88.3

    max

    90.4

  • CursorBench

    CursorBench · Cursor

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

    40.7

    $4.87

    medium

    43.3

    $6.94

    high

    44.7

    $9.00

    xhigh

    46.1

    $11.43

    max

    46.6

    $11.95

  • GPQA diamond

    Epoch AI Benchmarking Hub

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

    87.9

    max

    93.9

  • LMArena Text (style control)

    LMArena

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

    1549.4

    max

    1546.6

  • LMArena Text (style control)

    LMArena

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

    1532.0

    max

    1528.2

  • LMArena Text (style control)

    LMArena

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

    1470.8

    max

    1473.3

  • LMArena Text (style control)

    LMArena

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

    1493.7

    max

    1492.9

  • LMArena Text (style control)

    LMArena

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

    1503.9

    max

    1502.9

  • LMArena Text (style control)

    LMArena

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

    1536.3

    max

    1530.4

  • LMArena Text (style control)

    LMArena

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

    1507.4

    max

    1481.3

  • LMArena Text (style control)

    LMArena

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

    1487.8

    max

    1501.2

  • LMArena Text (style control)

    LMArena

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

    1515.3

    max

    1513.3

  • LMArena Text (style control)

    LMArena

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

    1513.7

    max

    1511.7

  • LMArena Text (style control)

    LMArena

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

    1485.5

    max

    1482.9

  • LMArena Text (style control)

    LMArena

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

    1463.6

    max

    1462.0

  • LMArena Text (style control)

    LMArena

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

    1499.8

    max

    1502.7

  • LMArena Text (style control)

    LMArena

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

    1509.8

    max

    1515.4

  • LMArena Text (style control)

    LMArena

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

    1530.6

    max

    1521.7

  • LMArena Text (style control)

    LMArena

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

    1503.7

    max

    1513.9

  • LMArena Text (style control)

    LMArena

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

    1521.8

    max

    1517.3

  • LMArena Text (style control)

    LMArena

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

    1480.9

    max

    1481.0

  • LMArena Text (style control)

    LMArena

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

    1496.3

    max

    1494.1

  • LMArena Text (style control)

    LMArena

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

    1505.8

    max

    1507.5

  • LMArena Text (style control)

    LMArena

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

    1481.5

    max

    1481.3

  • LMArena Text (style control)

    LMArena

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

    1501.6

    max

    1501.6

  • LMArena Text (style control)

    LMArena

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

    1521.3

    max

    1514.7

  • LMArena Text (style control)

    LMArena

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

    1485.8

    max

    1486.1

  • LMArena Text (style control)

    LMArena

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

    1479.4

    max

    1479.9

  • LMArena Text (style control)

    LMArena

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

    1489.7

    max

    1489.0

  • LMArena Text (style control)

    LMArena

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

    1501.6

    max

    1500.8

  • LMArena Text (style control)

    LMArena

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

    1496.5

    max

    1492.3

  • LMArena Text (style control)

    LMArena

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

    1477.5

    max

    1490.2

  • LMArena Text

    LMArena

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

    1570.2

    max

    1574.5

  • LMArena Text

    LMArena

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

    1530.8

    max

    1532.0

  • LMArena Text

    LMArena

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

    1486.0

    max

    1494.0

  • LMArena Text

    LMArena

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

    1502.5

    max

    1507.1

  • LMArena Text

    LMArena

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

    1518.2

    max

    1524.0

  • LMArena Text

    LMArena

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

    1552.1

    max

    1552.3

  • LMArena Text

    LMArena

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

    1519.9

    max

    1498.3

  • LMArena Text

    LMArena

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

    1502.0

    max

    1519.9

  • LMArena Text

    LMArena

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

    1522.2

    max

    1525.3

  • LMArena Text

    LMArena

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

    1518.7

    max

    1522.3

  • LMArena Text

    LMArena

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

    1497.1

    max

    1499.1

  • LMArena Text

    LMArena

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

    1477.0

    max

    1480.5

  • LMArena Text

    LMArena

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

    1521.5

    max

    1527.8

  • LMArena Text

    LMArena

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

    1520.8

    max

    1532.8

  • LMArena Text

    LMArena

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

    1542.9

    max

    1538.5

  • LMArena Text

    LMArena

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

    1509.4

    max

    1526.6

  • LMArena Text

    LMArena

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

    1525.9

    max

    1526.5

  • LMArena Text

    LMArena

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

    1493.8

    max

    1498.8

  • LMArena Text

    LMArena

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

    1515.2

    max

    1517.8

  • LMArena Text

    LMArena

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

    1514.1

    max

    1520.7

  • LMArena Text

    LMArena

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

    1502.1

    max

    1504.3

  • LMArena Text

    LMArena

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

    1511.9

    max

    1516.3

  • LMArena Text

    LMArena

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

    1532.2

    max

    1530.7

  • LMArena Text

    LMArena

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

    1493.6

    max

    1498.8

  • LMArena Text

    LMArena

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

    1495.0

    max

    1500.0

  • LMArena Text

    LMArena

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

    1502.2

    max

    1506.5

  • LMArena Text

    LMArena

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

    1517.6

    max

    1521.9

  • LMArena Text

    LMArena

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

    1505.8

    max

    1506.9

  • LMArena Text

    LMArena

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

    1505.5

    max

    1522.2

  • LMArena WebDev

    LMArena

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

    1656.5

    max

    1691.1

  • LMArena WebDev

    LMArena

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

    1656.5

    max

    1691.1

  • LMArena WebDev

    LMArena

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

    1650.2

    max

    1679.3

  • LMArena WebDev

    LMArena

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

    1678.1

    max

    1713.1

Closest head-to-head rivals

Direct pairwise win rates for Opus 5 across shared benchmark boards.