List your company

MLLM Safety Evaluation on OOD-MMSafe Constitution Mode 1.0 (test)

1.84RAGemma-3-12B
0.46720.82361.181.5364Mar 10, 2026

Evaluation results

MethodLinks
Gemma-3-12BModel Type=Open-source2026.031.842.91.826.61.960
GPT-5.1Model Type=Closed-source2026.031.827.71.837.51.980
Gemini-3-ProModel Type=Closed-source2026.031.799.21.88.61.940
Gemini-3-Pro-ThinkModel Type=Closed-source2026.031.712.71.7131.910.2
Qwen3VL-32BModel Type=Open-source2026.031.5817.61.6217.61.850.2
Qwen3VL-4BModel Type=Open-source2026.031.5416.71.6515.41.811.1
Doubao-1.6Model Type=Closed-source2026.031.42241.4822.41.780.7
InternVL-3.5-38BModel Type=Open-source2026.031.417.61.7211.61.654
Qwen3VL-8BModel Type=Open-source2026.031.3328.11.4225.71.760.9
o4-MiniModel Type=Closed-source2026.031.2831.91.3529.51.730.9
InternVL-3.5-8BModel Type=Open-source2026.030.7649.21.0543.41.272.6
LLaVA-1.5-7BModel Type=Open-source2026.030.5960.90.8352.10.9520.7
QwenVL2.5-7BModel Type=Open-source2026.030.5264.40.666.81.242.4