Medical Visual Question Answering on OmniMed OOD (test)
82.8AccuracyMeissa
Evaluation Results
| Method | Links | |
|---|---|---|
| MeissaFramework=Framework II, Inference Mode=Agent-based, Agent System=Ophiuchus2026.03 | 82.8 | |
| Ophiuchus-7BFramework=Framework II, Inference Mode=Agent-based, Agent System=Ophiuchus2026.03 | 78.6 | |
| Gemini-3-flashFramework=Framework II, Inference Mode=Direct Inference2026.03 | 78 | |
| Gemini-3-flashFramework=Framework II, Inference Mode=Agent-based, Agent System=Ophiuchus2026.03 | 77.4 | |
| Gemini-2.5-proFramework=Framework II, Inference Mode=Agent-based, Agent System=Ophiuchus2026.03 | 76.9 | |
| GPT-5Framework=Framework II, Inference Mode=Direct Inference2026.03 | 75.4 | |
| o3Framework=Framework II, Inference Mode=Agent-based, Agent System=Ophiuchus2026.03 | 73.7 | |
| Qwen3-VL-4BFramework=Framework II, Inference Mode=Direct Inference, Training Protocol=SFT2026.03 | 67.6 | |
| Qwen3-VL-4BFramework=Framework II, Inference Mode=Agent-based, Agent System=Ophiuchus2026.03 | 38.1 |