Agentic medical interaction on AgentClinic MedQA
65.8Accuracyo3
Evaluation results
| Method | Links | |
|---|---|---|
| o3Model category=Large Models, Patient agent=GPT-4o2025.07 | 65.8 | |
| Gemini 2.5 ProModel category=Large Models, Patient agent=GPT-4o2025.07 | 58.3 | |
| DeepSeek R1Model category=Large Models, Patient agent=GPT-4o2025.07 | 58.1 | |
| MedGemma 27BModel category=Small Models, Patient agent=GPT-4o2025.07 | 56.2 | |
| Human physicianModel category=Human metric, Patient agent=GPT-42025.07 | 54 | |
| Gemma 3 27BModel category=Small Models, Patient agent=GPT-4o2025.07 | 50.7 |