Multi-Turn Medical Dialogue on MedicalExam
74.33AccuracyMEDIQ
Evaluation Results
| Method | Links | |
|---|---|---|
| MEDIQModel=Gemini-2.5-Pro, Method Type=Prompt2026.02 | 74.33 | |
| ATPO (U1 + U2)Model=Qwen3-8B, Method Type=SFT+RL2026.02 | 65.87 | |
| ATPO (U1)Model=Qwen3-8B, Method Type=SFT+RL2026.02 | 65.52 | |
| TreePOModel=Qwen3-8B, Method Type=SFT+RL2026.02 | 65.33 | |
| MEDIQModel=GPT-4o, Method Type=Prompt2026.02 | 64 | |
| GRPOModel=Qwen3-8B, Method Type=SFT+RL2026.02 | 60.93 | |
| ATPO (U1 + U2)Model=Qwen3-4B, Method Type=SFT+RL2026.02 | 59.73 | |
| PPO (MDP)Model=Qwen3-8B, Method Type=SFT+RL2026.02 | 59.2 | |
| PPO (H-MDP)Model=Qwen3-8B, Method Type=SFT+RL2026.02 | 59.07 | |
| ATPO (U1)Model=Qwen3-4B, Method Type=SFT+RL2026.02 | 56.8 | |
| TreePOModel=Qwen3-4B, Method Type=SFT+RL2026.02 | 56.13 | |
| SFTModel=Qwen3-8B, Method Type=SFT2026.02 | 55.87 | |
| GRPOModel=Qwen3-4B, Method Type=SFT+RL2026.02 | 53.87 | |
| PPO (H-MDP)Model=Qwen3-4B, Method Type=SFT+RL2026.02 | 52.4 | |
| DirectModel=Qwen3-8B, Method Type=Prompt2026.02 | 52.4 | |
| MEDIQModel=Qwen3-8B, Method Type=Prompt2026.02 | 51.87 | |
| DFTModel=Qwen3-8B, Method Type=SFT2026.02 | 51.86 | |
| PPO (MDP)Model=Qwen3-4B, Method Type=SFT+RL2026.02 | 50.13 | |
| SFTModel=Qwen3-4B, Method Type=SFT2026.02 | 48.93 | |
| DirectModel=Qwen3-4B, Method Type=Prompt2026.02 | 48.13 | |
| MEDIQModel=Qwen3-4B, Method Type=Prompt2026.02 | 45.87 | |
| ATPO (U1)Model=Qwen3-1.7B, Method Type=SFT+RL2026.02 | 45.73 | |
| TreePOModel=Qwen3-1.7B, Method Type=SFT+RL2026.02 | 43.33 | |
| ATPO (U1 + U2)Model=Qwen3-1.7B, Method Type=SFT+RL2026.02 | 43.2 | |
| DFTModel=Qwen3-4B, Method Type=SFT2026.02 | 43.07 | |
| GRPOModel=Qwen3-1.7B, Method Type=SFT+RL2026.02 | 42.93 | |
| PPO (MDP)Model=Qwen3-1.7B, Method Type=SFT+RL2026.02 | 39.33 | |
| PPO (H-MDP)Model=Qwen3-1.7B, Method Type=SFT+RL2026.02 | 39.33 | |
| DirectModel=Qwen3-1.7B, Method Type=Prompt2026.02 | 35.07 | |
| MEDIQModel=Qwen3-1.7B, Method Type=Prompt2026.02 | 34 | |
| SFTModel=Qwen3-1.7B, Method Type=SFT2026.02 | 32.27 | |
| DFTModel=Qwen3-1.7B, Method Type=SFT2026.02 | 29.07 |