Close-Ended QA on SpineBench
89.46Accuracy (Text)SpineGPT
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| SpineGPTSize=7B2025.10 | 89.46 | 84.46 | 87.89 | 87.44 | |
| GPT5-miniSize=-2025.10 | 89.12 | 80.83 | 85.83 | 87.01 | |
| Gemini-2.5-ProSize=-2025.10 | 88.44 | 88.6 | 88.5 | 89.23 | |
| GPT5Size=-2025.10 | 87.41 | 79.97 | 84.46 | 85.54 | |
| O3Size=-2025.10 | 86.73 | 82.38 | 85.01 | 85.36 | |
| GPT-4oSize=-2025.10 | 86.73 | 81.7 | 84.74 | 81.6 | |
| GLM-4.5VSize=21B2025.10 | 85.71 | 81.35 | 83.98 | 83.26 | |
| Qwen2.5-VL-72BSize=72B2025.10 | 84.69 | 79.79 | 82.75 | 79.88 | |
| Gemini-2.5-FlashSize=-2025.10 | 83.67 | 80.83 | 82.55 | 83.93 | |
| Medgemma-27BSize=27B2025.10 | 83.33 | 80.83 | 82.34 | 76.66 | |
| Linshu-32BSize=32B2025.10 | 81.29 | 76.68 | 79.47 | 77.3 | |
| Claude4Size=-2025.10 | 79.59 | 79.79 | 79.67 | 80.28 | |
| HuatuoGPT-7BSize=7B2025.10 | 75.85 | 80.83 | 77.82 | 74.21 | |
| Qwen2.5VL-7BSize=7B2025.10 | 75.51 | 74.09 | 74.95 | 64.74 |