Mathematical Reasoning on AIME 22-24 (Pass@1, #Tokens)
45.1Pass@1Dynasor
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| DynasorModel=R1-7B2025.12 | 45.1 | 10,314 | |
| VanillaModel=R1-7B2025.12 | 44.4 | 13,709 | |
| TIPModel=R1-7B2025.12 | 44.4 | 11,112 | |
| CRESTModel=R1-7B2025.12 | 44.4 | 9,488 | |
| SEALModel=R1-7B2025.12 | 42.22 | 12,448 | |
| Soft-ThinkingModel=R1-7B2025.12 | 35.6 | 12,551 | |
| TIPModel=R1-1.5B2025.12 | 24.4 | 14,157 | |
| DynasorModel=R1-1.5B2025.12 | 24.12 | 15,337 | |
| SEALModel=R1-1.5B2025.12 | 22.2 | 14,207 | |
| CRESTModel=R1-1.5B2025.12 | 20 | 13,388 | |
| VanillaModel=R1-1.5B2025.12 | 17.8 | 17,034 | |
| Soft-ThinkingModel=R1-1.5B2025.12 | 12.2 | 18,418 |