Medical Report Generation on MIMIC-CXR (Metrics: ROUGE-L, CIDEr, SemScore)
31.4ROUGE-LLingshu-7B + Med-Scout
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Lingshu-7B + Med-ScoutModel Category=Medical MLLMs, Post-training strategy=Med-Scout2026.01 | 31.4 | 105.2 | 30.3 | |
| Lingshu-7BModel Category=Medical MLLMs2026.01 | 30.9 | 104.9 | 29.7 | |
| MedGemma-4B-ITModel Category=Medical MLLMs2026.01 | 27.4 | 83.8 | 30.1 | |
| HuatuoGPT-Vision-7B + Med-ScoutModel Category=Medical MLLMs, Post-training strategy=Med-Scout2026.01 | 25.7 | 79 | 25.8 | |
| GEMINI-3-FLASHModel Category=Proprietary Models2026.01 | 24.9 | 90 | 29.8 | |
| Qwen3-VL-8B-Instruct + Med-ScoutModel Category=General-purpose MLLMs, Post-training strategy=Med-Scout2026.01 | 23.8 | 68.1 | 21 | |
| HuatuoGPT-Vision-7BModel Category=Medical MLLMs2026.01 | 23.6 | 75.6 | 24.6 | |
| Qwen3-VL-4B-Instruct + Med-ScoutModel Category=General-purpose MLLMs, Post-training strategy=Med-Scout2026.01 | 23.4 | 65.2 | 21.5 | |
| Qwen2.5-VL-7B-InstructModel Category=General-purpose MLLMs2026.01 | 22.9 | 63.9 | 18.6 | |
| Qwen3-VL-4B-InstructModel Category=General-purpose MLLMs2026.01 | 21.8 | 60.9 | 19.8 | |
| Qwen2.5-VL-3B-InstructModel Category=General-purpose MLLMs2026.01 | 21.4 | 57.7 | 19.4 | |
| Qwen3-VL-8B-InstructModel Category=General-purpose MLLMs2026.01 | 21.3 | 64.8 | 19.6 | |
| InternVL3-8BModel Category=General-purpose MLLMs2026.01 | 20.7 | 64.3 | 21.1 | |
| LLaVA-Med-7BModel Category=Medical MLLMs2026.01 | 16.4 | 49.2 | 16.7 | |
| GPT-5Model Category=Proprietary Models2026.01 | 14.7 | 89.4 | 25.2 |