Acropolis

Grok 3 Mini Beta

xAIGrade B. Evidence grade B of A to D: good.ProprietaryReleased 9 Apr 2025Consistent across benchmarks
Compare with peers

Domain profile

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

  • General text (2 boards): 54 (scale 0 to 100)General text (2 boards)54
  • Math (1 boards): 37 (scale 0 to 100)Math (1 boards)37
  • Science (1 boards): 53 (scale 0 to 100)Science (1 boards)53

Benchmark appearances

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

BenchmarkRankScoreCostStatusObserved
FrontierMath-2025-02-28-Private

Epoch AI Benchmarking Hub · high

#66 / 101
5.9%[3.2% to 8.6%]UnknownStandard10 Apr 2025
FrontierMath-2025-02-28-Private

Epoch AI Benchmarking Hub · low

#81 / 101
2.8%[0.9% to 4.6%]UnknownStandard10 Apr 2025
LMArena Text

LMArena · Japanese · Default

#138 / 288
1,342[1,317 to 1,367]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Japanese · Default

#151 / 288
1,334[1,309 to 1,359]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Spanish · Default

#169 / 304
1,381[1,351 to 1,410]UnknownStandard11 Nov 2026
GPQA diamond

Epoch AI Benchmarking Hub · low

#170 / 338
76.3%[70.3% to 82.2%]UnknownStandard10 Apr 2025
LMArena Text

LMArena · Korean · Default

#173 / 289
1,305[1,275 to 1,335]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Polish · Default

#173 / 243
1,352[1,337 to 1,366]UnknownStandard11 Nov 2026
GPQA diamond

Epoch AI Benchmarking Hub · high

#178 / 338
75.5%[70.1% to 80.9%]UnknownStandard26 May 2025
LMArena Text

LMArena · German · Default

#180 / 320
1,347[1,322 to 1,372]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Creative writing · Default

#183 / 419
1,342[1,330 to 1,353]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Expert · Default

#187 / 371
1,367[1,349 to 1,384]UnknownStandard11 Nov 2026
LMArena Text

LMArena · French · Default

#189 / 301
1,355[1,318 to 1,392]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry legal and government · Default

#189 / 393
1,372[1,356 to 1,388]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Spanish · Default

#189 / 304
1,371[1,341 to 1,400]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Russian · Default

#192 / 385
1,353[1,339 to 1,367]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Korean · Default

#192 / 289
1,289[1,259 to 1,319]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · German · Default

#193 / 320
1,334[1,308 to 1,359]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry entertainment and sports and media · Default

#194 / 419
1,332[1,322 to 1,341]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry writing and literature and language · Default

#194 / 420
1,345[1,336 to 1,354]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Math · Default

#194 / 404
1,372[1,358 to 1,387]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry life and physical and social science · Default

#195 / 419
1,374[1,364 to 1,384]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Instruction following · Default

#196 / 421
1,346[1,338 to 1,354]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Longer query · Default

#196 / 399
1,357[1,348 to 1,366]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Non English · Default

#197 / 421
1,348[1,342 to 1,355]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry medicine and healthcare · Default

#198 / 391
1,366[1,349 to 1,382]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Overall · Default

#198 / 421
1,364[1,359 to 1,369]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Polish · Default

#198 / 243
1,336[1,322 to 1,350]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Exclude ties · Default

#199 / 421
1,329[1,322 to 1,336]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Hard prompts · Default

#200 / 421
1,364[1,357 to 1,370]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry business and management and financial operations · Default

#200 / 414
1,347[1,337 to 1,357]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Math · Default

#200 / 404
1,370[1,356 to 1,384]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Hard prompts English · Default

#201 / 419
1,375[1,367 to 1,384]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Creative writing · Default

#202 / 419
1,336[1,324 to 1,347]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Chinese · Default

#203 / 395
1,378[1,361 to 1,394]UnknownStandard11 Nov 2026
LMArena Text

LMArena · English · Default

#203 / 421
1,376[1,370 to 1,383]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Multi turn · Default

#205 / 419
1,351[1,341 to 1,361]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Expert · Default

#205 / 371
1,383[1,366 to 1,401]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · French · Default

#205 / 301
1,356[1,319 to 1,393]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry software and IT services · Default

#208 / 421
1,373[1,366 to 1,381]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Instruction following · Default

#208 / 421
1,353[1,345 to 1,361]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry mathematical · Default

#210 / 402
1,360[1,345 to 1,375]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Longer query · Default

#211 / 399
1,367[1,357 to 1,376]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Russian · Default

#211 / 385
1,351[1,337 to 1,365]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Exclude ties · Default

#214 / 421
1,321[1,314 to 1,328]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry writing and literature and language · Default

#214 / 420
1,341[1,332 to 1,350]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Non English · Default

#214 / 421
1,340[1,334 to 1,347]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Overall · Default

#214 / 421
1,358[1,353 to 1,363]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry entertainment and sports and media · Default

#215 / 419
1,325[1,315 to 1,334]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry mathematical · Default

#215 / 402
1,362[1,347 to 1,377]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Coding · Default

#216 / 416
1,367[1,357 to 1,376]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry life and physical and social science · Default

#217 / 419
1,372[1,362 to 1,382]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Chinese · Default

#218 / 395
1,375[1,358 to 1,391]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Hard prompts · Default

#219 / 421
1,370[1,364 to 1,377]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Hard prompts English · Default

#220 / 419
1,383[1,374 to 1,392]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · English · Default

#221 / 421
1,372[1,365 to 1,379]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Multi turn · Default

#221 / 419
1,345[1,335 to 1,355]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry business and management and financial operations · Default

#224 / 414
1,349[1,339 to 1,360]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry legal and government · Default

#224 / 393
1,363[1,347 to 1,379]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry medicine and healthcare · Default

#226 / 391
1,363[1,347 to 1,379]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry software and IT services · Default

#230 / 421
1,381[1,373 to 1,388]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Coding · Default

#232 / 416
1,386[1,377 to 1,396]UnknownStandard11 Nov 2026

Reasoning effort scaling

Same benchmark and harness measured across multiple reasoning effort settings.

  • FrontierMath-2025-02-28-Private

    Epoch AI Benchmarking Hub

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

    2.8

    high

    5.9

  • GPQA diamond

    Epoch AI Benchmarking Hub

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

    76.3

    high

    75.5

Closest head-to-head rivals

Direct pairwise win rates for Grok 3 Mini Beta across shared benchmark boards.