Medical Visual Question Answering on MIMIC-CXR
84.68AccuracyMedQwen
Evaluation results
| Method | Links | |
|---|---|---|
| MedQwen2026.04 | 84.68 | |
| LLaVA-Med-1.52026.04 | 75.79 | |
| RadFM2026.04 | 69.3 | |
| MedVInT2026.04 | 66.06 | |
| LLaVA-v1.62026.04 | 63.7 | |
| Med-Flamingo2026.04 | 61.27 | |
| Gemini-1.52026.04 | 61.02 | |
| GPT-4o2026.04 | 60.61 | |
| Qwen-2.5-VL2026.04 | 60.43 | |
| InternVL-22026.04 | 59.47 | |
| miniGPT-Med2026.04 | 53.92 | |
| T2R-R3Scale=4B2026.06 | 28.04 | |
| Frozen TeacherScale=4B2026.06 | 25.85 | |
| T2R-R2Scale=4B2026.06 | 25.61 | |
| T2R-R3Scale=2B2026.06 | 23.87 | |
| Judge RewardScale=4B2026.06 | 23.09 | |
| RLVR OnlyScale=4B2026.06 | 22.67 | |
| T2R-R2Scale=2B2026.06 | 22.17 | |
| Frozen TeacherScale=2B2026.06 | 20.88 | |
| T2R-R1Scale=2B2026.06 | 20.55 | |
| Base ModelScale=4B2026.06 | 20.45 | |
| T2R-R1Scale=4B2026.06 | 19.82 | |
| Judge RewardScale=2B2026.06 | 17.53 | |
| Base ModelScale=2B2026.06 | 12.98 | |
| RLVR OnlyScale=2B2026.06 | 11.85 |