Multi-choice Visual Question Answering on PMC-MI-Bench
90AccuracyLingshu-7B
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Lingshu-7BParameters=7B2025.11 | 90 | 90.8 | 90.8 | 91.8 | |
| M3LLM-8BParameters=8B2025.11 | 90 | 89.9 | 91.2 | 89.8 | |
| HealthGPT-14BParameters=14B2025.11 | 88 | 87.5 | 89.7 | 87.4 | |
| QWen2.5-VL-7BParameters=7B2025.11 | 84 | 84 | 86.3 | 85 | |
| MedGemma-27BParameters=27B2025.11 | 82 | 81 | 80.5 | 83.1 | |
| InternVL3-8BParameters=8B2025.11 | 78 | 78 | 80.9 | 79.2 | |
| HuatuoGPT-Vision-7BParameters=7B2025.11 | 74 | 69.7 | 71.9 | 75.6 | |
| LLaVA-NeXT-7BParameters=7B2025.11 | 70 | 70 | 74.4 | 72.3 | |
| LLaVA-7BParameters=7B2025.11 | 66 | 68 | 67.6 | 68.8 | |
| LLaVA-Med-7BParameters=7B2025.11 | 46 | 47.9 | 45 | 51.1 |