Mathematical Reasoning on HMMT 2024
23.5Pass@1 AccuracyOPD
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| OPDStudent-Teacher Pair=DeepSeek-R1-Distill-Qwen-1.5B / JustRL-DeepSeek-1.5B2026.05 | 23.5 | 6.9 | 0 | — | |
| PRUNE-OPD (top-p)Student-Teacher Pair=DeepSeek-R1-Distill-Qwen-1.5B / JustRL-DeepSeek-1.5B2026.05 | 23.3 | 4.3 | 37.7 | — | |
| OPD (Truncate 4k)Student-Teacher Pair=DeepSeek-R1-Distill-Qwen-1.5B / JustRL-DeepSeek-1.5B2026.05 | 22.7 | 4.8 | 30.4 | — | |
| PRUNE-OPD (overlap)Student-Teacher Pair=DeepSeek-R1-Distill-Qwen-1.5B / JustRL-DeepSeek-1.5B2026.05 | 22.7 | 4.1 | 40.6 | — | |
| PRUNE-OPD (top-p)Student-Teacher Pair=DeepSeek-R1-Distill-Qwen-1.5B / DeepSeek-R1-Distill-Qwen-7B2026.05 | 14.6 | 4.8 | 61.6 | — | |
| PRUNE-OPD (overlap)Student-Teacher Pair=DeepSeek-R1-Distill-Qwen-1.5B / DeepSeek-R1-Distill-Qwen-7B2026.05 | 14.6 | 4 | 68 | — | |
| OPDStudent-Teacher Pair=DeepSeek-R1-Distill-Qwen-1.5B / DeepSeek-R1-Distill-Qwen-7B2026.05 | 14.4 | 12.5 | 0 | — | |
| OPD (Truncate 4k)Student-Teacher Pair=DeepSeek-R1-Distill-Qwen-1.5B / DeepSeek-R1-Distill-Qwen-7B2026.05 | 14.2 | 6.6 | 47.2 | — | |
| PRUNE-OPD (overlap)Student-Teacher Pair=Qwen3-4B-Base / Qwen3-4B (Non-thinking)2026.05 | 6.7 | 13.5 | 52.6 | — | |
| OPDStudent-Teacher Pair=Qwen3-4B-Base / Qwen3-4B (Non-thinking)2026.05 | 5.4 | 28.5 | 0 | — | |
| OPD (Truncate 4k)Student-Teacher Pair=Qwen3-4B-Base / Qwen3-4B (Non-thinking)2026.05 | 4.8 | 22.3 | 21.8 | — | |
| PRUNE-OPD (top-p)Student-Teacher Pair=Qwen3-4B-Base / Qwen3-4B (Non-thinking)2026.05 | 4.8 | 20 | 29.8 | — | |
| PRUNE-OPD (overlap)Student-Teacher Pair=Qwen3-1.7B-Base / Qwen3-4B (Non-thinking)2026.05 | 2.9 | 10.3 | 37.6 | — | |
| OPD (Truncate 4k)Student-Teacher Pair=Qwen3-1.7B-Base / Qwen3-4B (Non-thinking)2026.05 | 2.3 | 12.9 | 21.8 | — | |
| OPDStudent-Teacher Pair=Qwen3-1.7B-Base / Qwen3-4B (Non-thinking)2026.05 | 2.1 | 16.5 | 0 | — | |
| PRUNE-OPD (top-p)Student-Teacher Pair=Qwen3-1.7B-Base / Qwen3-4B (Non-thinking)2026.05 | 1.9 | 11.5 | 30.3 | — | |
| OPDTeacher-Student Pair=DeepSeek-R1-Distill-Qwen-7B / Skywork-OR1-7B, Time=17.5h, Red.=0%2026.05 | — | — | — | 33.5 | |
| OPD (Truncate 4k)Teacher-Student Pair=DeepSeek-R1-Distill-Qwen-7B / Skywork-OR1-7B, Time=11.3h, Red.=35.4%2026.05 | — | — | — | 31 | |
| PRUNE-OPD (overlap)Teacher-Student Pair=DeepSeek-R1-Distill-Qwen-7B / Skywork-OR1-7B, Time=17.0h, Red.=2.9%2026.05 | — | — | — | 33.1 |