Acropolis

Qwen4 Max

AlibabaGrade C. Evidence grade C of A to D: thin.ProprietaryConsistent across benchmarks
Compare with peers

Domain profile

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

  • Agentic (9 boards): 73 (scale 0 to 100)Agentic (9 boards)73
  • Coding (4 boards): 70 (scale 0 to 100)Coding (4 boards)70
  • Finance (4 boards): 74 (scale 0 to 100)Finance (4 boards)74
  • Legal (3 boards): 73 (scale 0 to 100)Legal (3 boards)73

Benchmark appearances

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

BenchmarkRankScoreCostStatusObserved
Legal Agent Benchmark (HLAB-Vals)

Vals AI benchmarks · xhigh

#7 / 38Tied #1
21.3%[14.7% to 27.8%]$12.87Standard15 Nov 2026
Finance Agent

Vals AI benchmarks · xhigh

#8 / 38Tied #1
66.7%[62.7% to 70.6%]$1.74Standard15 Nov 2026
Tax Agent

Vals AI benchmarks · xhigh

#10 / 38
40.4%[33.4% to 47.3%]$1.03Standard15 Nov 2026
Vals Index

Vals AI benchmarks · xhigh

#11 / 38
63.2%[61.2% to 65.2%]$9.33Standard15 Nov 2026
Vibe Code Bench

Vals AI benchmarks · xhigh · OpenHands

#11 / 38
86.7%[82.2% to 91.1%]$41.40Standard15 Nov 2026
Terminal-Bench 4

Vals AI benchmarks · xhigh

#13 / 38
48.5%[46.8% to 50.2%]$3.90Standard15 Nov 2026
Code Migration

Vals AI benchmarks · xhigh

#15 / 38
57.3%[49.7% to 64.8%]$10.11Standard15 Nov 2026
EMB

Vals AI benchmarks · xhigh

#17 / 38
70.8%[66.2% to 75.5%]$1.00Standard15 Nov 2026
Legal Research

Vals AI benchmarks · xhigh

#20 / 38
41.3%[34.6% to 48.1%]$2.59Standard15 Nov 2026

Closest head-to-head rivals

Direct pairwise win rates for Qwen4 Max across shared benchmark boards.