Medical Question Answering / Clinical Simulation on MedQA (test)
97.9AccuracyGemini-3-flash
Evaluation Results
| Method | Links | |
|---|---|---|
| Gemini-3-flashFramework=Framework IV, Inference Mode=Agent-based, Agent System=AgentClinic2026.03 | 97.9 | |
| GPT-4VFramework=Framework III, Inference Mode=Agent-based, Agent System=MDAgents2026.03 | 88.7 | |
| Gemini-3-flashFramework=Framework III, Inference Mode=Direct Inference2026.03 | 75.5 | |
| Gemini-3-flashFramework=Framework III, Inference Mode=Agent-based, Agent System=MDAgents2026.03 | 75.5 | |
| GPT-4VFramework=Framework III, Inference Mode=Direct Inference2026.03 | 75 | |
| Qwen3-VL-4BFramework=Framework III, Inference Mode=Agent-based, Agent System=MDAgents2026.03 | 59.8 | |
| MeissaFramework=Framework III, Inference Mode=Agent-based, Agent System=MDAgents2026.03 | 57.2 | |
| Llama-3.3-70BFramework=Framework IV, Inference Mode=Agent-based, Agent System=AgentClinic2026.03 | 54.7 | |
| GPT-4oFramework=Framework IV, Inference Mode=Agent-based, Agent System=AgentClinic2026.03 | 52.8 | |
| Qwen3-VL-4BFramework=Framework III, Inference Mode=Direct Inference, Training Protocol=SFT2026.03 | 51.8 | |
| Qwen3-VL-4BFramework=Framework IV, Inference Mode=Agent-based, Agent System=AgentClinic2026.03 | 50.5 | |
| MeissaFramework=Framework IV, Inference Mode=Agent-based, Agent System=AgentClinic2026.03 | 49.5 | |
| Qwen2.5-72BFramework=Framework IV, Inference Mode=Agent-based, Agent System=AgentClinic2026.03 | 38.7 |