Multi-turn Agent Task on ALFWorld
85.16Success RateATOD
Evaluation results
| Method | Links | ||
|---|---|---|---|
| ATODStudent Model=Qwen3-4B, Teacher Model=Qwen3-30B-A3B GRPO2026.06 | 85.16 | 16.75 | |
| ATODStudent Model=Qwen3-0.6B, Teacher Model=Qwen3-4B GRPO2026.06 | 82.81 | 15.84 | |
| ATODStudent Model=Qwen3-1.7B, Teacher Model=Qwen3-4B GRPO2026.06 | 80.47 | 22.69 | |
| GRPOStudent Model=None, Teacher Model=Qwen3-30B-A3B GRPO2026.06 | 80.47 | 23.53 | |
| OPDStudent Model=Qwen3-4B, Teacher Model=Qwen3-30B-A3B GRPO2026.06 | 80.47 | 21.12 | |
| SODStudent Model=Qwen3-4B, Teacher Model=Qwen3-30B-A3B GRPO2026.06 | 79.69 | 19.02 | |
| SDARStudent Model=Qwen3-4B, Teacher Model=None2026.06 | 77.34 | 20.85 | |
| GRPOStudent Model=None, Teacher Model=Qwen3-4B GRPO2026.06 | 76.56 | 18.95 | |
| OPDStudent Model=Qwen3-0.6B, Teacher Model=Qwen3-4B GRPO2026.06 | 76.56 | 18.48 | |
| SODStudent Model=Qwen3-0.6B, Teacher Model=Qwen3-4B GRPO2026.06 | 76.56 | 22.9 | |
| GRPOStudent Model=Qwen3-4B, Teacher Model=None2026.06 | 76.56 | 18.95 | |
| TCODStudent Model=Qwen3-4B, Teacher Model=Qwen3-30B-A3B GRPO2026.06 | 76.56 | 23.45 | |
| SODStudent Model=Qwen3-1.7B, Teacher Model=Qwen3-4B GRPO2026.06 | 75.78 | 27.95 | |
| TCODStudent Model=Qwen3-0.6B, Teacher Model=Qwen3-4B GRPO2026.06 | 74.22 | 20.15 | |
| OPDStudent Model=Qwen3-1.7B, Teacher Model=Qwen3-4B GRPO2026.06 | 72.66 | 16.41 | |
| TCODStudent Model=Qwen3-1.7B, Teacher Model=Qwen3-4B GRPO2026.06 | 71.88 | 18.88 | |
| SDARStudent Model=Qwen3-1.7B, Teacher Model=None2026.06 | 39.84 | 37.56 | |
| GRPOStudent Model=Qwen3-1.7B, Teacher Model=None2026.06 | 31.25 | 40.38 | |
| GRPOStudent Model=Qwen3-0.6B, Teacher Model=None2026.06 | 30.47 | 44.13 | |
| SDARStudent Model=Qwen3-0.6B, Teacher Model=None2026.06 | 28.12 | 45.27 | |
| VanillaStudent Model=Qwen3-4B, Teacher Model=None2026.06 | 24.22 | 42.43 | |
| VanillaStudent Model=Qwen3-1.7B, Teacher Model=None2026.06 | 10.16 | 47.91 | |
| VanillaStudent Model=Qwen3-0.6B, Teacher Model=None2026.06 | 0.78 | 50 |