List your company

Detection on CareQA-Vision Adversarial (test)

82Base AccuracyAloe-Vision
22.7238.1153.568.89Jun 25, 2026

Evaluation results

MethodLinks
Aloe-VisionSize=72B, Stage=S22026.0682315754.333.1
Qwen3-VL-8B-Instruct2026.068119.519.735.716.4
Qwen3-VL-30B-A3B-Instruct2026.068120.837.134.326.2
GLM-4.5V2026.06803.812.620.214.4
Qwen2-VLSize=72B2026.0678.46.35.720.58.7
LingshuSize=7B2026.0678.14.78.622.317.8
LingshuSize=32B2026.0677.77.812.634.822.7
Aloe-VisionSize=7B, Stage=S22026.0676.975.358.568.658
Kimi-VL-A3B-Instruct2026.0675.84.27.41021.1
Hulu-MedSize=32B2026.0673.338.221.815.3
Aloe-VisionSize=72B, Stage=S12026.0672.542.919.79.4
GPT-5.22026.0670.914.35.39.78.7
Aloe-VisionSize=7B, Stage=S12026.0669.166.410.326.334.4
Hulu-MedSize=7B2026.0668.91.80.1526.3
Qwen2-VLSize=7B2026.0665.116.58.517.622.6
MiMo-VLSize=7B2026.0663.90.91.58.92.3
Chiron-o1Size=8B2026.0657.419.28.633.324.4
HuatuoGPT-VisionSize=34B2026.0655.925.62.36.336.6
HuatuoGPT-VisionSize=7B2026.0651.925.50.75.241.6
Random2026.0625——25—