Interactive Medical Diagnosis on AgentClinic MedQA
53.9ICRQwen2.5-72B
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Qwen2.5-72BInteractive Strategy=REFINE2026.01 | 53.9 | 64.5 | |
| GPT-5-miniInteractive Strategy=REFINE2026.01 | 45.5 | 68 | |
| Qwen2.5-72BInteractive Strategy=SC2026.01 | 43.6 | 62 | |
| Qwen2.5-7BInteractive Strategy=Baseline2026.01 | 43.5 | 43 | |
| Qwen2.5-72BInteractive Strategy=Baseline2026.01 | 41.7 | 54 | |
| GPT-5-miniInteractive Strategy=SC2026.01 | 40.2 | 69.5 | |
| Qwen2.5-72BInteractive Strategy=ReAct2026.01 | 39.5 | 59 | |
| Qwen2.5-7BInteractive Strategy=REFINE2026.01 | 38.6 | 45.5 | |
| Qwen2.5-7BInteractive Strategy=SC2026.01 | 36 | 44.5 | |
| Qwen2.5-7BInteractive Strategy=ReAct2026.01 | 35.8 | 45 | |
| GPT-5-miniInteractive Strategy=ReAct2026.01 | 33.2 | 67 | |
| GPT-5-miniInteractive Strategy=Baseline2026.01 | 33 | 61 |