| Attribute | Value |
|---|---|
| Base Model | Qwen3-1.7B |
| Teacher Model | SOD-GRPO_teacher-4B |
| Training Pipeline | Cold-Start SFT → SOD (Step-wise On-policy Distillation) |
| Parameters | 1.7B |
| Model | Description |
|---|---|
| SOD-0.6B | SOD-distilled 0.6B student |
| SOD-1.7B | SOD-distilled 1.7B student (this model) |
| SOD-GRPO_teacher-4B | GRPO-trained 4B teacher model |
| Method | AIME 2024 | AIME 2025 | GPQA-Diamond | LiveCodeBench-v6 | Average |
|---|---|---|---|---|---|
| Vanilla | 9.90 | 8.96 | 26.80 | 22.73 | 17.10 |
| SFT | 26.77 | 22.40 | 29.85 | 24.63 | 25.91 |
| GRPO | 25.63 | 21.67 | 33.55 | 20.70 | 25.39 |
| OPD | 43.86 | 37.04 | 31.73 | 32.45 | 36.27 |
| OPSD_gt | 33.85 | 24.69 | 35.02 | 22.73 | 29.07 |
| OPSD_hint | 34.42 | 21.43 | 33.46 | 23.12 | 28.11 |
| SOD (This Model) | 50.83 | 41.72 | 38.72 | 40.63 | 42.98 |
| Method | AIME 2024 | AIME 2025 | GPQA-Diamond | LiveCodeBench-v6 | Average |
|---|---|---|---|---|---|
| GRPO | 67.60 | 60.42 | 55.19 | 63.13 | 61.59 |
1@article{zhong2026sod,
2 title={SOD: Step-wise On-policy Distillation for Small Language Model Agents},
3 author={Zhong, Qiyong and Zheng, Mao and Song, Mingyang and Lin, Xin and Sun, Jie and Jiang, Houcheng and Wang, Xiang and Fang, Junfeng},
4 journal={arXiv preprint arXiv:2605.07725},
5 year={2026}
6}