Code Generation on MBPP (MBPP Pass@1)
65.1MBPP Pass@1Trajectory Replay
Evaluation Results
| Method | Links | |
|---|---|---|
| Trajectory ReplayDonor=QWEN3-VL-8B, Target=GLM-4.6V, Type=DENSE->REAS2026.01 | 65.1 | |
| BaseDonor=QWEN3-VL-8B, Target=GLM-4.6V, Type=DENSE->REAS2026.01 | 65 | |
| BaseDonor=QWEN3-VL-2B, Target=MINISTRAL-3-14B, Type=CROSS-FAM2026.01 | 62.1 | |
| BaseDonor=MINISTRAL-3-3B, Target=MINISTRAL-3-14B, Type=IN-FAM2026.01 | 62.1 | |
| Trajectory ReplayDonor=MINISTRAL-3-3B, Target=MINISTRAL-3-14B, Type=IN-FAM2026.01 | 62 | |
| Trajectory ReplayDonor=QWEN3-VL-2B, Target=MINISTRAL-3-14B, Type=CROSS-FAM2026.01 | 61.8 | |
| BaseDonor=MINISTRAL-3-3B, Target=QWEN3-VL-8B, Type=REVERSE2026.01 | 58.4 | |
| BaseDonor=QWEN3-VL-2B, Target=QWEN3-VL-8B, Type=IN-FAM2026.01 | 58.4 | |
| Trajectory ReplayDonor=QWEN3-VL-2B, Target=QWEN3-VL-8B, Type=IN-FAM2026.01 | 58.3 | |
| Trajectory ReplayDonor=MINISTRAL-3-3B, Target=QWEN3-VL-8B, Type=REVERSE2026.01 | 58.1 | |
| BaseDonor=QWEN3-VL-8B, Target=GPT-OSS-20B, Type=DENSE->MOE2026.01 | 55.2 | |
| Trajectory ReplayDonor=QWEN3-VL-8B, Target=GPT-OSS-20B, Type=DENSE->MOE2026.01 | 55 | |
| BaseDonor=QWEN3-VL-8B, Target=MINISTRAL-3-3B, Type=CROSS-FAM2026.01 | 48 | |
| Trajectory ReplayDonor=QWEN3-VL-8B, Target=MINISTRAL-3-3B, Type=CROSS-FAM2026.01 | 47.9 | |
| BaseDonor=MINISTRAL-3-14B, Target=QWEN3-VL-2B, Type=REVERSE2026.01 | 31 | |
| Trajectory ReplayDonor=MINISTRAL-3-14B, Target=QWEN3-VL-2B, Type=REVERSE2026.01 | 30.8 |