Mathematical Reasoning on AIME25 (Acc, Tok, EPR, Avg. EPR)
0.423AccuracyVanilla
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| VanillaModel=DS-7B2026.01 | 0.423 | 14,319 | — | — | |
| EntrocutModel=DS-7B2026.01 | 0.381 | 9,555 | 2 | 2.4 | |
| DEERModel=DS-7B2026.01 | 0.377 | 9,586 | 1.8 | 1.5 | |
| TIPModel=DS-7B2026.01 | 0.329 | 10,138 | 0.8 | 1.3 | |
| NOWAITModel=DS-7B2026.01 | 0.3 | 7,449 | — | — | |
| VanillaModel=DS-1.5B2026.01 | 0.254 | 15,099 | — | — | |
| EntrocutModel=DS-1.5B2026.01 | 0.238 | 7,912 | 4 | 12.4 | |
| TIPModel=DS-1.5B2026.01 | 0.229 | 9,160 | 2.1 | 2.9 | |
| DEERModel=DS-1.5B2026.01 | 0.216 | 8,807 | 1.5 | 1.7 | |
| NOWAITModel=DS-1.5B2026.01 | 0.188 | 7,772 | — | — |