Classification on CareQA-Vision Adversarial (test)
78.8Base AccuracyLingshu
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| LingshuSize=7B2026.06 | 78.8 | 1.2 | 18.8 | 45.4 | |
| GPT-5.22026.06 | 70.2 | 6.4 | 10.9 | 38.5 | |
| GLM-4.5V2026.06 | 70.2 | 0.3 | 8.6 | 21 | |
| LingshuSize=32B2026.06 | 68.9 | 2.7 | 29.8 | 56.9 | |
| Hulu-MedSize=32B2026.06 | 68.4 | 1.3 | 4.1 | 35.4 | |
| Qwen3-VL-8B-Instruct2026.06 | 67.6 | 0.3 | 8.6 | 28.2 | |
| Qwen3-VL-30B-A3B-Instruct2026.06 | 66.8 | 0.1 | 13.2 | 28.9 | |
| Aloe-VisionSize=72B, Stage=S22026.06 | 66.6 | 16.2 | 51 | 63.5 | |
| Aloe-VisionSize=7B, Stage=S22026.06 | 65.3 | 13.7 | 43.8 | 50.2 | |
| Aloe-VisionSize=72B, Stage=S12026.06 | 63 | 3.5 | 8.2 | 75 | |
| Hulu-MedSize=7B2026.06 | 61.2 | 1.3 | 0 | 18.6 | |
| Aloe-VisionSize=7B, Stage=S12026.06 | 59.6 | 2.3 | 15 | 41.6 | |
| Qwen2-VLSize=72B2026.06 | 59.4 | 1 | 5.8 | 35.6 | |
| HuatuoGPT-VisionSize=34B2026.06 | 59.1 | 22.7 | 10.1 | 14.9 | |
| HuatuoGPT-VisionSize=7B2026.06 | 57.3 | 21.4 | 7.2 | 30.7 | |
| Kimi-VL-A3B-Instruct2026.06 | 55.6 | 0 | 0.6 | 11.5 | |
| MiMo-VLSize=7B2026.06 | 54.6 | 0.3 | 2.2 | 6.4 | |
| Qwen2-VLSize=7B2026.06 | 51.6 | 0.2 | 2.4 | 10.1 | |
| Chiron-o1Size=8B2026.06 | 48.9 | 5.8 | 8.6 | 54.6 | |
| Random2026.06 | 41.67 | — | 41.67 | — |