Mathematical Reasoning on AIME24 (Acc, Tok, EPR, Avg. EPR)
55.8AccVanilla
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| VanillaModel=DS-7B2026.01 | 55.8 | 12,775 | — | — | |
| EntrocutModel=DS-7B2026.01 | 51.7 | 9,663 | 1.5 | 2.4 | |
| TIPModel=DS-7B2026.01 | 51.3 | 9,919 | 1.2 | 1.3 | |
| DEERModel=DS-7B2026.01 | 49.2 | 9,138 | 1.1 | 1.5 | |
| NOWAITModel=DS-7B2026.01 | 46.3 | 7,921 | — | — | |
| VanillaModel=DS-1.5B2026.01 | 29.2 | 15,591 | — | — | |
| EntrocutModel=DS-1.5B2026.01 | 28.8 | 8,295 | 17.5 | 12.4 | |
| TIPModel=DS-1.5B2026.01 | 27.9 | 11,458 | 3.1 | 2.9 | |
| DEERModel=DS-1.5B2026.01 | 27.5 | 8,695 | 3.9 | 1.7 | |
| NOWAITModel=DS-1.5B2026.01 | 22.1 | 8,196 | — | — |