Medical Visual Question Answering on SLAKE (Acc, Recall, Rouge)
71.55AccuracyOurs (Med-Evo)
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Ours (Med-Evo)base model=MedVLM-R12026.03 | 71.55 | 30.03 | 39.66 | |
| TTRLbase model=MedVLM-R12026.03 | 71.27 | 26.2 | 36.72 | |
| EN-INFbase model=MedVLM-R12026.03 | 68.73 | 25.21 | 33.5 | |
| TTRVbase model=MedVLM-R12026.03 | 68.45 | 25.64 | 33.74 | |
| DeepEyes-v2Model Type=Think with Images Methods2026.05 | 66.2 | — | — | |
| MAESTROModel Type=Ours2026.05 | 66.2 | — | — | |
| GLM-4.6VModel Type=Open-Source & Baselines2026.05 | 63.1 | — | — | |
| ThymeModel Type=Think with Images Methods2026.05 | 62.6 | — | — | |
| GPT-5Model Type=Closed-Source2026.05 | 61.8 | — | — | |
| Base Modelbase model=MedVLM-R12026.03 | 61.41 | 17.99 | 26.82 | |
| VTOOL-R1Model Type=Think with Images Methods2026.05 | 60.7 | — | — | |
| PixelReasonerModel Type=Think with Images Methods2026.05 | 59.8 | — | — | |
| Kimi-K2.5Model Type=Open-Source & Baselines2026.05 | 59.6 | — | — | |
| DeepEyesModel Type=Think with Images Methods2026.05 | 58.7 | — | — | |
| GPT-4oModel Type=Closed-Source2026.05 | 58.5 | — | — | |
| Visual-ARFTModel Type=Think with Images Methods2026.05 | 58.5 | — | — | |
| VTS-VModel Type=Think with Images Methods2026.05 | 57.9 | — | — | |
| Qwen3-VL-32BModel Type=Open-Source & Baselines2026.05 | 57.6 | — | — | |
| Gemini-2.5-ProModel Type=Closed-Source2026.05 | 56.8 | — | — | |
| VisionReasonerModel Type=Think with Images Methods2026.05 | 56.8 | — | — | |
| Gemini-2.5-FlashModel Type=Closed-Source2026.05 | 56 | — | — | |
| Direct AnsweringModel Type=Open-Source & Baselines2026.05 | 56 | — | — | |
| Untrained ModelModel Type=Open-Source & Baselines2026.05 | 54.3 | — | — | |
| MathCoder-VLModel Type=Think with Images Methods2026.05 | 54 | — | — | |
| Chain-of-FocusModel Type=Think with Images Methods2026.05 | 48.2 | — | — |