Visual Question Answering on OPENMEDREASON-Bench
78.51AccuracyQwen2.5-VL-7B + OPENMEDREASON
Evaluation results
| Method | Links | |
|---|---|---|
| Qwen2.5-VL-7B + OPENMEDREASONMedical Samples=450K2026.06 | 78.51 | |
| Lingshu 7BMedical Samples=5M2026.06 | 63.2 | |
| MedVL-Thinker 7BMedical Samples=200k2026.06 | 52.03 | |
| OctoMed 7BMedical Samples=8M2026.06 | 51.16 | |
| MedGemma 27BMedical Samples=30M2026.06 | 48.22 | |
| QoQ-Med-VL 7BMedical Samples=2.6M2026.06 | 47.86 | |
| Qwen2.5-VL-7BMedical Samples=N/A2026.06 | 47.11 | |
| MedGemma 1.5 4BMedical Samples=30M2026.06 | 46.19 |