Mathematical Reasoning on MATH (pv, best, Ac/Y metrics)
87.6Ac (alpha=0.10)R-Q8B
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| R-Q8BModel=Qwen3-1.7B2026.05 | 87.6 | 72.8 | 69.2 | 77.6 | 92.5 | 68.7 | |
| R-L8BModel=Qwen3-1.7B2026.05 | 87.4 | 71.8 | 66.2 | 74.9 | 92.5 | 62.9 | |
| MVModel=Qwen3-1.7B2026.05 | 87.3 | 71.2 | — | 71.2 | 92.4 | 57.9 | |
| Ppl-StdModel=Qwen3-1.7B2026.05 | 86.6 | 70.8 | 63.2 | 73.6 | 92.6 | 62.1 | |
| J-Q30BModel=Qwen3-1.7B2026.05 | 86.3 | 69.8 | 57.5 | 72 | 91.7 | 59.4 | |
| PplModel=Qwen3-1.7B2026.05 | 86.2 | 69.8 | 63.5 | 72.7 | 91.6 | 61.4 | |
| J-G26BModel=Qwen3-1.7B2026.05 | 86 | 70 | 52 | 72.1 | 91.4 | 59 | |
| J-R1Q14BModel=Qwen3-1.7B2026.05 | 85.5 | 69.1 | 54.7 | 70.5 | 91 | 56.2 | |
| OracleModel=Qwen3-1.7B2026.05 | 85.4 | — | 85.4 | 100 | 85.4 | 100 | |
| R-Q8BModel=R1-Qwen-1.5B2026.05 | 82.2 | 64.8 | 62.2 | 55.2 | 84.1 | 29.6 | |
| J-R1Q14BModel=R1-Qwen-1.5B2026.05 | 81 | 60.8 | 58.3 | 50.3 | 83.4 | 29.5 | |
| MVModel=R1-Qwen-1.5B2026.05 | 80.9 | 60.3 | — | 46.5 | 82.8 | 24 | |
| R-L8BModel=R1-Qwen-1.5B2026.05 | 80.9 | 62.3 | 57.5 | 50.1 | 83.3 | 27.1 | |
| J-G26BModel=R1-Qwen-1.5B2026.05 | 80.9 | 61.8 | 56.7 | 52.3 | 83.8 | 30 | |
| J-Q30BModel=R1-Qwen-1.5B2026.05 | 80.6 | 60.6 | 52.3 | 48.8 | 83.1 | 29 | |
| Ppl-StdModel=R1-Qwen-1.5B2026.05 | 79.4 | 59.6 | 53.2 | 46.5 | 82.6 | 27.7 | |
| PplModel=R1-Qwen-1.5B2026.05 | 77.8 | 57.2 | 49.8 | 42.9 | 81.5 | 25.1 | |
| OracleModel=R1-Qwen-1.5B2026.05 | 76.5 | — | 76.5 | 100 | 76.5 | 100 | |
| J-G26BModel=Gemma-4-E4B2026.05 | 70.3 | 58.7 | 54.2 | 31.1 | 72.3 | 7.1 | |
| R-Q8BModel=Gemma-4-E4B2026.05 | 70.2 | 60.9 | 57.7 | 32.2 | 64.6 | 7.5 | |
| J-R1Q14BModel=Gemma-4-E4B2026.05 | 69.2 | 58.6 | 51.5 | 30.7 | 77.7 | 6.9 | |
| R-L8BModel=Gemma-4-E4B2026.05 | 69.1 | 58.7 | 57.1 | 31.5 | 72.7 | 9.5 | |
| OracleModel=Gemma-4-E4B2026.05 | 68.5 | — | 68.5 | 100 | 68.5 | 100 | |
| MVModel=Gemma-4-E4B2026.05 | 68.5 | 58.3 | — | 21.1 | — | 0 | |
| PplModel=Gemma-4-E4B2026.05 | 67.9 | 51.1 | 46.8 | 29.1 | 84.1 | 6.1 | |
| Ppl-StdModel=Gemma-4-E4B2026.05 | 67.8 | 50.6 | 45.9 | 29.1 | 67.3 | 7.7 | |
| J-Q30BModel=Gemma-4-E4B2026.05 | 67.4 | 56.8 | 51.2 | 28.1 | 52.9 | 1.6 | |
| GreedyModel=Qwen3-1.7B2026.05 | 55.5 | — | — | 100 | 55.5 | 100 | |
| GreedyModel=Gemma-4-E4B2026.05 | 52.1 | — | — | 100 | 52.1 | 100 | |
| GreedyModel=R1-Qwen-1.5B2026.05 | 47.6 | — | — | 100 | 47.6 | 100 |