Mathematical Reasoning on NuminaMath (pass@1, pass@5)
45.96Pass@1DeepSeek-R1-Qwen-1.5B
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| DeepSeek-R1-Qwen-1.5BGroup Size (G)=162026.06 | 45.96 | 60.61 | |
| DeepSeek-R1-Qwen-1.5BGroup Size (G)=42026.06 | 44.95 | 60.11 | |
| DeepSeek-R1-Qwen-1.5BGroup Size (G)=Base2026.06 | 42.42 | 61.62 | |
| DeepSeek-7BGroup Size (G)=42026.06 | 17.57 | 35.35 | |
| Nemotron-4BGroup Size (G)=162026.06 | 17.37 | 32.32 | |
| DeepSeek-7BGroup Size (G)=Base2026.06 | 16.97 | 30.3 | |
| DeepSeek-7BGroup Size (G)=162026.06 | 15.96 | 28.79 | |
| Nemotron-4BGroup Size (G)=Base2026.06 | 14.55 | 33.33 | |
| Nemotron-4BGroup Size (G)=42026.06 | 14.55 | 30.81 |