Mathematical Reasoning on MATH500 (Avg@32, Pass@32)
89.8Average Score (Avg@32)SCOPE (Ours)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SCOPE (Ours)Teacher Model=SkyWork-OR1-Math-7B, Student Model=DeepSeek-R1-Distill-Qwen-1.5B2026.04 | 89.8 | 97.9 | |
| w/ OPDTeacher Model=SkyWork-OR1-Math-7B, Student Model=DeepSeek-R1-Distill-Qwen-1.5B2026.04 | 89 | 97.7 | |
| w/ GRPOTeacher Model=SkyWork-OR1-Math-7B, Student Model=DeepSeek-R1-Distill-Qwen-1.5B2026.04 | 87 | 96.7 | |
| R1-Distill-Qwen-1.5BTeacher Model=SkyWork-OR1-Math-7B, Student Model=DeepSeek-R1-Distill-Qwen-1.5B2026.04 | 84.6 | 97.3 | |
| w/ KDTeacher Model=SkyWork-OR1-Math-7B, Student Model=DeepSeek-R1-Distill-Qwen-1.5B2026.04 | 84.1 | 97.4 | |
| SCOPE (Ours)Teacher Model=Qwen3-8B-Instruct, Student Model=Qwen3-1.7B-Base2026.04 | 70.9 | 93 | |
| w/ OPDTeacher Model=Qwen3-8B-Instruct, Student Model=Qwen3-1.7B-Base2026.04 | 67.9 | 90.7 | |
| w/ GRPOTeacher Model=Qwen3-8B-Instruct, Student Model=Qwen3-1.7B-Base2026.04 | 62.9 | 89.8 | |
| Qwen3-1.7B-BaseTeacher Model=Qwen3-8B-Instruct, Student Model=Qwen3-1.7B-Base2026.04 | 54.8 | 88.6 | |
| w/ KDTeacher Model=Qwen3-8B-Instruct, Student Model=Qwen3-1.7B-Base2026.04 | 50.2 | 86.2 |