Mathematical Reasoning on AMC AIME (test)
67.4AccuracyPIR
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| PIRBackbone=DeepSeek-R1-32B2026.02 | 67.4 | 2,935.83 | |
| Specutive ThinkingBackbone=DeepSeek-R1-32B2026.02 | 64.6 | 4,941.73 | |
| QwQ-32BBackbone=QwQ-32B2026.02 | 63 | 4,600.75 | |
| L1-maxBackbone=DeepSeek-R1-32B2026.02 | 60.6 | 3,832.18 | |
| DeepSeek-R1-32BBackbone=DeepSeek-R1-32B2026.02 | 59.92 | 3,661.52 | |
| DeepSeek-R1-7BBackbone=DeepSeek-R1-7B2026.02 | 58.33 | 3,858.16 | |
| DeepSeek-R1-14BBackbone=DeepSeek-R1-14B2026.02 | 57.76 | 3,523.09 | |
| prompt to conciseBackbone=DeepSeek-R1-32B2026.02 | 54.5 | 3,286.95 | |
| s1-32BBackbone=DeepSeek-R1-32B2026.02 | 53.96 | 5,958.03 | |
| DEERBackbone=DeepSeek-R1-32B2026.02 | 51.93 | 4,112.46 | |
| SpecReasonBackbone=DeepSeek-R1-32B2026.02 | 43 | 6,354.78 | |
| SEALBackbone=DeepSeek-R1-32B2026.02 | 42.34 | 6,354.78 |