Acropolis

Haiku 4.5

AnthropicGrade A. Evidence grade A of A to D: strongest.ProprietaryReleased 15 Oct 2025Consistent across benchmarks
Compare with peers

Domain profile

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

  • Agentic (3 boards): 14 (scale 0 to 100)Agentic (3 boards)14
  • Coding (3 boards): 16 (scale 0 to 100)Coding (3 boards)16
  • Finance (1 boards): 14 (scale 0 to 100)Finance (1 boards)14
  • General text (5 boards): 25 (scale 0 to 100)General text (5 boards)25
  • Math (1 boards): 39 (scale 0 to 100)Math (1 boards)39
  • Science (2 boards): 37 (scale 0 to 100)Science (2 boards)37

Benchmark appearances

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

BenchmarkRankScoreCostStatusObserved
LMArena Document

LMArena · Overall · Default

#39 / 46
1,420[1,415 to 1,426]UnknownStandard9 Nov 2026
LMArena Document (style control)

LMArena · Overall · Default

#44 / 46
1,419[1,413 to 1,424]UnknownStandard9 Nov 2026
Vending-Bench 2

Epoch AI Benchmarking Hub · Default

#63 / 72
$458.89UnknownStandardNot stated
FrontierMath-2025-02-28-Private

Epoch AI Benchmarking Hub · Default

#65 / 101
5.9%[3.2% to 8.6%]UnknownStandard22 Oct 2025
FrontierMath-2025-02-28-Private

Epoch AI Benchmarking Hub · Default

#72 / 101
4.1%[1.8% to 6.4%]UnknownStandard16 Oct 2025
LMArena Text

LMArena · Expert · Default

#93 / 371
1,443[1,437 to 1,449]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Coding · Default

#99 / 416
1,453[1,450 to 1,457]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Instruction following · Default

#100 / 421
1,414[1,410 to 1,418]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Longer query · Default

#101 / 399
1,426[1,423 to 1,430]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Spanish · Default

#105 / 304
1,431[1,421 to 1,441]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Coding · Default

#113 / 416
1,482[1,478 to 1,486]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Spanish · Default

#114 / 304
1,422[1,412 to 1,432]UnknownStandard11 Nov 2026
Terminal Bench

Epoch AI Benchmarking Hub · Goose

#116 / 202
35.5%[29.8% to 41.2%]UnknownStandard15 Oct 2025
LMArena Text

LMArena · Hard prompts English · Default

#120 / 419
1,435[1,431 to 1,439]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Expert · Default

#120 / 371
1,452[1,446 to 1,458]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry software and IT services · Default

#120 / 421
1,465[1,461 to 1,468]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Longer query · Default

#120 / 399
1,436[1,432 to 1,439]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry mathematical · Default

#122 / 402
1,433[1,426 to 1,440]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry mathematical · Default

#123 / 402
1,429[1,422 to 1,437]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Hard prompts English · Default

#124 / 419
1,454[1,450 to 1,458]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry software and IT services · Default

#125 / 421
1,439[1,436 to 1,443]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Korean · Default

#125 / 289
1,365[1,351 to 1,378]UnknownStandard11 Nov 2026
LMArena WebDev

LMArena · Webdev react · Default

#125 / 140
1,325[1,319 to 1,331]UnknownStandard15 Nov 2026
LMArena Text (style control)

LMArena · Instruction following · Default

#128 / 421
1,415[1,411 to 1,418]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry writing and literature and language · Default

#130 / 420
1,399[1,395 to 1,403]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Multi turn · Default

#130 / 419
1,425[1,420 to 1,429]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · French · Default

#132 / 301
1,432[1,421 to 1,443]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Japanese · Default

#132 / 288
1,363[1,347 to 1,378]UnknownStandard11 Nov 2026
LMArena WebDev

LMArena · Overall · Default

#132 / 156
1,330[1,325 to 1,335]UnknownStandard15 Nov 2026
LMArena WebDev

LMArena · Webdev · Default

#132 / 156
1,330[1,325 to 1,335]UnknownStandard15 Nov 2026
LMArena Text (style control)

LMArena · English · Default

#133 / 421
1,431[1,428 to 1,435]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry business and management and financial operations · Default

#133 / 414
1,419[1,415 to 1,423]UnknownStandard11 Nov 2026
LMArena WebDev

LMArena · Webdev HTML · Default

#133 / 155
1,318[1,308 to 1,327]UnknownStandard15 Nov 2026
LMArena Text (style control)

LMArena · German · Default

#134 / 320
1,400[1,387 to 1,414]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry entertainment and sports and media · Default

#134 / 419
1,388[1,384 to 1,392]UnknownStandard11 Nov 2026
Terminal Bench

Epoch AI Benchmarking Hub · Mini-SWE-Agent

#134 / 202
29.8%[24.9% to 34.8%]UnknownStandard15 Oct 2025
LMArena Text (style control)

LMArena · Hard prompts · Default

#135 / 421
1,441[1,438 to 1,444]UnknownStandard11 Nov 2026
Terminal Bench

Epoch AI Benchmarking Hub · Mini-SWE-Agent

#135 / 202
29.8%[24.9% to 34.7%]UnknownStandard3 Nov 2025
LMArena Text

LMArena · Hard prompts · Default

#137 / 421
1,420[1,417 to 1,423]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Creative writing · Default

#138 / 419
1,387[1,382 to 1,392]UnknownStandard11 Nov 2026
Terminal Bench

Epoch AI Benchmarking Hub · Terminus 2

#139 / 202
28.3%[22.7% to 34%]UnknownStandard15 Oct 2025
Terminal Bench

Epoch AI Benchmarking Hub · Terminus 2

#140 / 202
28.3%[22.6% to 34%]UnknownStandard31 Oct 2025
LMArena Text

LMArena · Industry writing and literature and language · Default

#141 / 420
1,386[1,382 to 1,390]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Japanese · Default

#141 / 288
1,340[1,324 to 1,356]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Multi turn · Default

#141 / 419
1,409[1,405 to 1,414]UnknownStandard11 Nov 2026
Bug Hunt Bench

Bug Hunt Bench · Claude Code

#143 / 145
1.5$2.22Standard8 Oct 2026
LMArena Text

LMArena · Industry business and management and financial operations · Default

#143 / 414
1,399[1,395 to 1,404]UnknownStandard11 Nov 2026
Terminal Bench

Epoch AI Benchmarking Hub · Claude Code

#143 / 202
27.5%[22.1% to 33%]UnknownStandard15 Oct 2025
LMArena Text

LMArena · Creative writing · Default

#144 / 419
1,371[1,366 to 1,375]UnknownStandard11 Nov 2026
LMArena Text

LMArena · French · Default

#144 / 301
1,411[1,400 to 1,421]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry entertainment and sports and media · Default

#144 / 419
1,369[1,365 to 1,373]UnknownStandard11 Nov 2026
Terminal Bench

Epoch AI Benchmarking Hub · Claude Code

#144 / 202
27.5%[22% to 33%]UnknownStandard4 Nov 2025
LMArena Text (style control)

LMArena · Industry life and physical and social science · Default

#145 / 419
1,431[1,427 to 1,436]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Korean · Default

#147 / 289
1,344[1,331 to 1,358]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Exclude ties · Default

#147 / 421
1,402[1,398 to 1,405]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Overall · Default

#149 / 421
1,414[1,412 to 1,416]UnknownStandard11 Nov 2026
LMArena Text

LMArena · English · Default

#151 / 421
1,414[1,411 to 1,418]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry legal and government · Default

#153 / 393
1,416[1,410 to 1,422]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Russian · Default

#154 / 385
1,401[1,396 to 1,406]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Chinese · Default

#155 / 395
1,434[1,427 to 1,440]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Non English · Default

#155 / 421
1,394[1,391 to 1,397]UnknownStandard11 Nov 2026
LMArena Text

LMArena · German · Default

#157 / 320
1,373[1,360 to 1,386]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry legal and government · Default

#158 / 393
1,400[1,394 to 1,406]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry medicine and healthcare · Default

#162 / 391
1,425[1,418 to 1,432]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Math · Default

#162 / 404
1,402[1,395 to 1,409]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Exclude ties · Default

#164 / 421
1,376[1,373 to 1,379]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Math · Default

#164 / 404
1,397[1,391 to 1,404]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Overall · Default

#164 / 421
1,396[1,394 to 1,399]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Polish · Default

#164 / 243
1,371[1,360 to 1,381]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry medicine and healthcare · Default

#165 / 391
1,394[1,387 to 1,400]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry life and physical and social science · Default

#166 / 419
1,406[1,401 to 1,411]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Russian · Default

#167 / 385
1,381[1,375 to 1,386]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Chinese · Default

#171 / 395
1,418[1,412 to 1,425]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Non English · Default

#171 / 421
1,376[1,373 to 1,379]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Polish · Default

#182 / 243
1,343[1,333 to 1,354]UnknownStandard11 Nov 2026
Terminal Bench

Epoch AI Benchmarking Hub · OpenHands

#188 / 202
13.9%[8.6% to 19.3%]UnknownStandard15 Oct 2025
Terminal Bench

Epoch AI Benchmarking Hub · OpenHands

#189 / 202
13.9%[8.6% to 19.2%]UnknownStandard2 Nov 2025
GPQA diamond

Epoch AI Benchmarking Hub · Default

#199 / 338
71.2%[64.9% to 77.5%]UnknownStandard22 Oct 2025
ARC-AGI-2

Epoch AI Benchmarking Hub · Default

#214 / 299
4%$0.38StandardNot stated
ARC-AGI-2

Epoch AI Benchmarking Hub · Default

#225 / 299
2.8%$0.14StandardNot stated
GPQA diamond

Epoch AI Benchmarking Hub · Default

#230 / 338
60.5%[55% to 66%]UnknownStandard16 Oct 2025
ARC-AGI-2

Epoch AI Benchmarking Hub · Default

#247 / 299
1.7%$0.09StandardNot stated
ARC-AGI-2

Epoch AI Benchmarking Hub · Default

#259 / 299
1.3%$0.04StandardNot stated
ARC-AGI-2

Epoch AI Benchmarking Hub · Default

#259 / 299
1.3%$0.05StandardNot stated

Closest head-to-head rivals

Direct pairwise win rates for Haiku 4.5 across shared benchmark boards.