Medical Question Answering on MedXpertQA OOD Text-only
54.6Accuracy (OOD Text-only)o3
Evaluation Results
| Method | Links | |
|---|---|---|
| o3Model size category=Large Models2025.07 | 54.6 | |
| Gemini 2.5 ProModel size category=Large Models2025.07 | 43.1 | |
| Gemini 2.5 FlashModel size category=Large Models2025.07 | 36.2 | |
| MedGemma 27BModel size category=Small Models2025.07 | 25.7 | |
| Gemma 3 27BModel size category=Small Models2025.07 | 15.7 | |
| MedGemma 4BModel size category=Small Models2025.07 | 14.2 | |
| Gemma 3 4BModel size category=Small Models2025.07 | 11.6 |