Captioning on MRI-VQA image-level (test)
25.05AR-ScoreQwen3-VL-8B-SGMRIVQA
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Qwen3-VL-8B-SGMRIVQAParameters=8B, Training=Fine-tuned2026.04 | 25.05 | 59.45 | |
| Gemini-2.5-FlashModel Type=Proprietary2026.04 | 21.72 | 49.81 | |
| Qwen3-VL-8BModel Type=Open-Source, Parameters=8B2026.04 | 19.86 | 41.66 | |
| Qwen2.5-VL-7BModel Type=Open-Source, Parameters=7B2026.04 | 18.19 | 23.72 | |
| MedGemma-1.5Model Type=Domain-Specific Medical, Parameters=4B, Context=image-level only2026.04 | 17.24 | 32.61 | |
| GPT-4oModel Type=Proprietary2026.04 | 16.9 | 32.41 | |
| Eagle2.5-8BModel Type=Open-Source, Parameters=8B2026.04 | 15.41 | 31.31 | |
| InternVL2.5-8BModel Type=Open-Source, Parameters=8B2026.04 | 14.37 | 22 | |
| LLaVA-Video-7BModel Type=Open-Source, Parameters=7B2026.04 | 13.31 | 38.02 | |
| LLaVA-Med-v1.5Model Type=Domain-Specific Medical, Parameters=7B, Context=image-level only2026.04 | 12.2 | 31.18 | |
| Gemini-2.5-ProModel Type=Proprietary2026.04 | 11.91 | 19.17 |