Acropolis

Beam 1.0

ReflectionGrade A. Evidence grade A of A to D: strongest.Open weightsReleased 16 Oct 2026Consistent across benchmarks
Compare with peers

Domain profile

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

  • Agentic (10 boards): 12 (scale 0 to 100)Agentic (10 boards)12
  • Coding (6 boards): 12 (scale 0 to 100)Coding (6 boards)12
  • Finance (4 boards): 8 (scale 0 to 100)Finance (4 boards)8
  • General text (3 boards): 75 (scale 0 to 100)General text (3 boards)75
  • Legal (3 boards): 12 (scale 0 to 100)Legal (3 boards)12
  • Science (1 boards): 49 (scale 0 to 100)Science (1 boards)49

Benchmark appearances

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

BenchmarkRankScoreCostStatusObserved
Legal Agent Benchmark (HLAB-Vals)

Vals AI benchmarks · Default

#32 / 38
4.6%[1.1% to 8.1%]UnknownStandard15 Nov 2026
EMB

Vals AI benchmarks · Default

#33 / 38
64.7%[60.3% to 69.2%]UnknownStandard15 Nov 2026
Legal Research

Vals AI benchmarks · Default

#33 / 38
28.8%[22.7% to 35%]UnknownStandard15 Nov 2026
Vibe Code Bench

Vals AI benchmarks · OpenHands

#33 / 38
66.4%[58.5% to 74.4%]UnknownStandard15 Nov 2026
Terminal-Bench 4

Vals AI benchmarks · Default

#34 / 38
4%[1.4% to 6.7%]UnknownStandard15 Nov 2026
Code Migration

Vals AI benchmarks · Default

#35 / 38
22.6%[15.8% to 29.4%]UnknownStandard15 Nov 2026
Finance Agent

Vals AI benchmarks · Default

#35 / 38
51.8%[47.7% to 56%]UnknownStandard15 Nov 2026
Vals Index

Vals AI benchmarks · Default

#35 / 38
43.4%[41.3% to 45.5%]UnknownStandard15 Nov 2026
Tax Agent

Vals AI benchmarks · Default

#38 / 38
8.8%[5.1% to 12.5%]UnknownStandard15 Nov 2026
LMArena Agent

LMArena · Overall · Default

#51 / 64
-0.1[-0.1 to -0]UnknownStandard15 Nov 2026
LMArena Text (style control)

LMArena · Math · Default

#52 / 404
1,471[1,433 to 1,510]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry mathematical · Default

#56 / 402
1,475[1,436 to 1,514]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Math · Default

#58 / 404
1,459[1,421 to 1,498]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry mathematical · Default

#59 / 402
1,466[1,427 to 1,505]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Coding · Default

#63 / 416
1,475[1,456 to 1,493]UnknownStandard11 Nov 2026
LMArena Text

LMArena · French · Default

#65 / 301
1,460[1,420 to 1,500]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Coding · Default

#69 / 416
1,505[1,486 to 1,524]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Expert · Default

#71 / 371
1,464[1,434 to 1,493]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Expert · Default

#77 / 371
1,481[1,452 to 1,511]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Hard prompts English · Default

#78 / 419
1,453[1,434 to 1,473]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry software and IT services · Default

#82 / 421
1,461[1,445 to 1,477]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Instruction following · Default

#83 / 421
1,424[1,408 to 1,440]UnknownStandard11 Nov 2026
LMArena WebDev

LMArena · Webdev react · Default

#84 / 140
1,442[1,424 to 1,459]UnknownStandard15 Nov 2026
LMArena Text (style control)

LMArena · French · Default

#87 / 301
1,466[1,426 to 1,507]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Hard prompts English · Default

#88 / 419
1,472[1,453 to 1,492]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry software and IT services · Default

#90 / 421
1,484[1,468 to 1,500]UnknownStandard11 Nov 2026
LMArena WebDev

LMArena · Overall · Default

#91 / 156
1,436[1,423 to 1,450]UnknownStandard15 Nov 2026
LMArena WebDev

LMArena · Webdev · Default

#91 / 156
1,436[1,423 to 1,450]UnknownStandard15 Nov 2026
LMArena Text

LMArena · Hard prompts · Default

#92 / 421
1,442[1,430 to 1,454]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry business and management and financial operations · Default

#92 / 414
1,426[1,403 to 1,448]UnknownStandard11 Nov 2026
LMArena WebDev

LMArena · Webdev HTML · Default

#92 / 155
1,430[1,393 to 1,467]UnknownStandard15 Nov 2026
LMArena Text

LMArena · Industry entertainment and sports and media · Default

#95 / 419
1,401[1,382 to 1,419]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Instruction following · Default

#95 / 421
1,433[1,417 to 1,449]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Hard prompts · Default

#97 / 421
1,461[1,449 to 1,473]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Longer query · Default

#98 / 399
1,428[1,413 to 1,443]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry business and management and financial operations · Default

#102 / 414
1,441[1,418 to 1,463]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Non English · Default

#105 / 421
1,412[1,400 to 1,424]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry legal and government · Default

#106 / 393
1,433[1,400 to 1,465]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · English · Default

#106 / 421
1,446[1,430 to 1,461]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry entertainment and sports and media · Default

#106 / 419
1,404[1,386 to 1,423]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Exclude ties · Default

#107 / 421
1,415[1,402 to 1,428]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Overall · Default

#109 / 421
1,424[1,415 to 1,434]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Exclude ties · Default

#109 / 421
1,429[1,416 to 1,442]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Overall · Default

#109 / 421
1,434[1,425 to 1,444]UnknownStandard11 Nov 2026
LMArena Text

LMArena · English · Default

#110 / 421
1,435[1,420 to 1,450]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Non English · Default

#111 / 421
1,421[1,409 to 1,433]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry writing and literature and language · Default

#117 / 420
1,398[1,381 to 1,416]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Multi turn · Default

#117 / 419
1,421[1,397 to 1,445]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Multi turn · Default

#117 / 419
1,432[1,408 to 1,456]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Longer query · Default

#118 / 399
1,438[1,423 to 1,453]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry legal and government · Default

#120 / 393
1,437[1,405 to 1,470]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Creative writing · Default

#121 / 419
1,388[1,368 to 1,409]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Russian · Default

#122 / 385
1,419[1,395 to 1,444]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry writing and literature and language · Default

#124 / 420
1,405[1,387 to 1,422]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Industry life and physical and social science · Default

#127 / 419
1,443[1,419 to 1,467]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Russian · Default

#128 / 385
1,405[1,380 to 1,430]UnknownStandard11 Nov 2026
LMArena Text

LMArena · Industry life and physical and social science · Default

#130 / 419
1,429[1,405 to 1,453]UnknownStandard11 Nov 2026
LMArena Text (style control)

LMArena · Creative writing · Default

#137 / 419
1,388[1,367 to 1,409]UnknownStandard11 Nov 2026
ARC-AGI-2

Epoch AI Benchmarking Hub · xhigh

#160 / 299
23.5%UnknownStandardNot stated
ARC-AGI-2

Epoch AI Benchmarking Hub · high

#164 / 299
20%UnknownStandardNot stated
LMArena Text

LMArena · Industry medicine and healthcare · Default

#187 / 391
1,376[1,341 to 1,411]UnknownStandard11 Nov 2026
ARC-AGI-2

Epoch AI Benchmarking Hub · medium

#195 / 299
6.4%UnknownStandardNot stated
LMArena Text (style control)

LMArena · Industry medicine and healthcare · Default

#197 / 391
1,394[1,359 to 1,429]UnknownStandard11 Nov 2026
ARC-AGI-2

Epoch AI Benchmarking Hub · low

#226 / 299
2.8%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.
    low

    2.8

    medium

    6.4

    high

    20.0

    xhigh

    23.5

Closest head-to-head rivals

Direct pairwise win rates for Beam 1.0 across shared benchmark boards.