List your company

Jailbreak Attack Evaluation on HarmfulQA

16ASRDeepSeek V3.1
14.723.47532.2541.025Jan 8, 2026

Evaluation results

MethodLinks
DeepSeek V3.1Model=DeepSeek V3.12026.0116
Grok 3 MiniModel=Grok 3 Mini2026.0117.5
Gemini 2.5 FlashModel=Gemini 2.5 Flash2026.0118
Qwen2.5 7BModel=Qwen2.5 7B2026.0121
GPT-4o-miniModel=GPT-4o-mini2026.0122.5
Mixtral 8×7BModel=Mixtral 8×7B2026.0148.5