ReportQA on MIMIC-CXR
61.26QAScoreGPT-5.4
Evaluation results
| Method | Links | |
|---|---|---|
| GPT-5.4Category=Proprietary VLMs2026.06 | 61.26 | |
| Gemini 3.1 ProCategory=Proprietary VLMs2026.06 | 50.55 | |
| Claude Opus 4.6Category=Proprietary VLMs2026.06 | 49.95 | |
| Hulu-Med-32BCategory=Open-source VLMs (medical), Parameters=32B2026.06 | 46.87 | |
| Qwen3.5-27BCategory=Open-source VLMs (general), Parameters=27B2026.06 | 46.57 | |
| InternVL3.5-38BCategory=Open-source VLMs (general), Parameters=38B2026.06 | 40.72 | |
| MedGemma 1.5 (4B)Category=Open-source VLMs (medical), Parameters=4B2026.06 | 33.94 | |
| Lingshu-32BCategory=Open-source VLMs (medical), Parameters=32B2026.06 | 24.83 | |
| RadFMCategory=Open-source VLMs (medical)2026.06 | 16.32 |