Mathematical Reasoning on AIME N=30, 3 seeds 2026
93.3AccuracyOpus 120k
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Opus 120kN=30, Seeds=32026.05 | 93.3 | 87.8 | 0.273 | 0.25 | |
| DASEW=8, N=30, Seeds=32026.05 | 90 | 83.3 | 0.145 | — | |
| DASEW=2, N=30, Seeds=32026.05 | 85.6 | 77.8 | 0.041 | 0.125 | |
| S1 ConsensusN=30, Seeds=32026.05 | 76.7 | 67.8 | 0.014 | 0.001 | |
| Opus 4.6 Std.N=30, Seeds=32026.05 | 60 | 50 | 0.057 | 0.001 |