Medical Diagnosis on MedAction 300 Hard
82Diag. Acc.GPT-5.4
Evaluation Results
| Method | Links | |
|---|---|---|
| GPT-5.4Model Source=Closed-Source2026.05 | 82 | |
| II-Medical-8B SFTModel Source=Open-Source, Model Parameters=8B, Training Stage=SFT2026.05 | 69 | |
| OpenAI-o3-miniModel Source=Closed-Source2026.05 | 67 | |
| Baichuan-M3Model Source=Open-Source, Model Parameters=235B2026.05 | 66 | |
| Gemini 2.5 Flash-LiteModel Source=Closed-Source2026.05 | 63 | |
| MedGemmaModel Source=Open-Source, Model Parameters=27B2026.05 | 63 | |
| Qwen-QwQModel Source=Open-Source, Model Parameters=32B2026.05 | 58 | |
| Baichuan-M1Model Source=Open-Source, Model Parameters=14B2026.05 | 56 | |
| II-Medical-8BModel Source=Open-Source, Model Parameters=8B, Training Stage=Base2026.05 | 46 |