Acropolis

GPT-4.5 Preview

OpenAIGrade B. Evidence grade B of A to D: good.ProprietaryWithin-domain variance
Compare with peers

Domain profile

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

  • General text (4 boards): 44 (scale 0 to 100)General text (4 boards)44
  • Math (1 boards): 16 (scale 0 to 100)Math (1 boards)16
  • Science (3 boards): 16 (scale 0 to 100)Science (3 boards)16
  • Vision (2 boards): 50 (scale 0 to 100)Vision (2 boards)50

Benchmark appearances

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

BenchmarkRankScoreCostStatusObserved
LMArena Text (style control)

LMArena · German · Default

#36 / 320
1,477[1,445 to 1,509]UnknownStandard11 Nov 2026
HLE

Epoch AI Benchmarking Hub · Default

#49 / 56
5.4%[3.7% to 7.2%]UnknownStandardNot stated
LMArena Text

LMArena · German · Default

#49 / 320
1,457[1,424 to 1,489]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Japanese · Default

#53 / 288
1,439[1,401 to 1,478]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Japanese · Default

#60 / 288
1,416[1,378 to 1,454]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry entertainment and sports and media · Default

#63 / 419
1,435[1,424 to 1,447]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry writing and literature and language · Default

#63 / 420
1,447[1,438 to 1,457]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Creative writing · Default

#65 / 419
1,436[1,424 to 1,448]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry legal and government · Default

#65 / 393
1,469[1,446 to 1,491]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Korean · Default

#71 / 289
1,414[1,371 to 1,458]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Multi turn · Default

#73 / 419
1,466[1,452 to 1,479]UnknownStandard11 Nov 2026
LMArena Vision (style control)

LMArena · Overall · Default

#78 / 167
1,225[1,214 to 1,237]UnknownStandard9 Nov 2026
LMArena Text (style control)

LMArena · Non English · Default

#80 / 421
1,438[1,430 to 1,446]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Exclude ties · Default

#81 / 421
1,452[1,444 to 1,461]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Korean · Default

#82 / 289
1,392[1,349 to 1,436]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Multi turn · Default

#82 / 419
1,444[1,431 to 1,457]UnknownStandard11 Nov 2026
LMArena Vision (style control)

LMArena · English · Default

#82 / 167
1,218[1,202 to 1,235]UnknownStandard9 Nov 2026
LMArena Text (style control)

LMArena · Instruction following · Default

#85 / 421
1,437[1,428 to 1,445]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Russian · Default

#85 / 385
1,447[1,432 to 1,463]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry writing and literature and language · Default

#89 / 420
1,412[1,402 to 1,421]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Overall · Default

#91 / 421
1,445[1,439 to 1,450]UnknownStandard11 Nov 2026
LMArena Vision (style control)

LMArena · Chinese · Default

#94 / 135
1,207[1,147 to 1,268]UnknownStandard9 Nov 2026
LMArena Vision

LMArena · Overall · Default

#97 / 167
1,195[1,183 to 1,207]UnknownStandard9 Nov 2026
LMArena Text

LMArena · Industry entertainment and sports and media · Default

#98 / 419
1,399[1,387 to 1,411]UnknownStandard11 Nov 2026
LMArena Vision

LMArena · Chinese · Default

#100 / 135
1,176[1,113 to 1,239]UnknownStandard9 Nov 2026
LMArena Vision

LMArena · English · Default

#101 / 167
1,190[1,174 to 1,207]UnknownStandard9 Nov 2026
LMArena Text

LMArena · Non English · Default

#102 / 421
1,413[1,405 to 1,421]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · English · Default

#103 / 421
1,447[1,440 to 1,454]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Russian · Default

#104 / 385
1,419[1,403 to 1,434]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Longer query · Default

#104 / 399
1,443[1,430 to 1,457]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry legal and government · Default

#105 / 393
1,434[1,411 to 1,456]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Creative writing · Default

#113 / 419
1,394[1,382 to 1,406]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry medicine and healthcare · Default

#113 / 391
1,450[1,426 to 1,474]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Exclude ties · Default

#114 / 421
1,411[1,402 to 1,419]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry life and physical and social science · Default

#118 / 419
1,448[1,437 to 1,460]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · French · Default

#120 / 301
1,443[1,387 to 1,500]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Instruction following · Default

#123 / 421
1,404[1,395 to 1,412]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry business and management and financial operations · Default

#123 / 414
1,425[1,410 to 1,441]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Overall · Default

#126 / 421
1,417[1,412 to 1,423]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry software and IT services · Default

#130 / 421
1,458[1,448 to 1,468]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Chinese · Default

#131 / 395
1,456[1,436 to 1,476]UnknownStandard11 Nov 2026
LMArena Text

LMArena · French · Default

#134 / 301
1,418[1,362 to 1,474]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Hard prompts · Default

#137 / 421
1,440[1,430 to 1,450]UnknownStandard11 Nov 2026
LMArena Text

LMArena · English · Default

#139 / 421
1,420[1,413 to 1,427]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Longer query · Default

#140 / 399
1,406[1,392 to 1,419]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Math · Default

#141 / 404
1,412[1,397 to 1,427]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Hard prompts English · Default

#144 / 419
1,442[1,430 to 1,455]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry life and physical and social science · Default

#146 / 419
1,420[1,409 to 1,432]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry medicine and healthcare · Default

#149 / 391
1,411[1,387 to 1,435]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Coding · Default

#150 / 416
1,458[1,445 to 1,471]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Expert · Default

#153 / 371
1,429[1,405 to 1,452]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Math · Default

#153 / 404
1,407[1,392 to 1,423]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry mathematical · Default

#155 / 402
1,411[1,395 to 1,426]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry mathematical · Default

#157 / 402
1,412[1,396 to 1,427]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Hard prompts · Default

#165 / 421
1,403[1,393 to 1,413]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry business and management and financial operations · Default

#165 / 414
1,386[1,370 to 1,401]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Hard prompts English · Default

#167 / 419
1,407[1,394 to 1,419]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry software and IT services · Default

#168 / 421
1,412[1,402 to 1,422]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Chinese · Default

#169 / 395
1,421[1,402 to 1,441]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Expert · Default

#171 / 371
1,394[1,371 to 1,417]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Coding · Default

#183 / 416
1,396[1,383 to 1,409]UnknownStandard11 Nov 2026
GPQA diamond

Epoch AI Benchmarking Hub · Default

#206 / 338
68.7%[62.2% to 75.2%]UnknownStandard28 Feb 2025
ARC-AGI-2

Epoch AI Benchmarking Hub · Default

#274 / 299
0.8%$2.10StandardNot stated

Closest head-to-head rivals

Direct pairwise win rates for GPT-4.5 Preview across shared benchmark boards.