List your company

Safety Evaluation on HarmBench Contextual Trajectory Evaluation Multi-turn

96.1SFRSTAR
-0.30824.72149.7574.779Mar 15, 2026

Evaluation results

MethodLinks
STARTarget Model=Gemini 2.0-Flash, Access Type=Closed-Source Models2026.0396.1
STARTarget Model=GPT 4o, Access Type=Closed-Source Models2026.0394.5
X-teamingTarget Model=GPT 4o, Access Type=Closed-Source Models2026.0394.3
STARTarget Model=Llama 3-8B-IT, Access Type=Open-Weight Models2026.0389
X-teamingTarget Model=Gemini 2.0-Flash, Access Type=Closed-Source Models2026.0387.4
X-teamingTarget Model=Llama 3-8B-IT, Access Type=Open-Weight Models2026.0385.5
STARTarget Model=Llama 3-70B-IT, Access Type=Open-Weight Models2026.0385.5
ActorAttackTarget Model=Llama 3-70B-IT, Access Type=Open-Weight Models2026.0385
X-teamingTarget Model=Llama 3-70B-IT, Access Type=Open-Weight Models2026.0384.9
ActorAttackTarget Model=GPT 4o, Access Type=Closed-Source Models2026.0384.5
RACETarget Model=GPT 4o, Access Type=Closed-Source Models2026.0382.8
ActorAttackTarget Model=Llama 3-8B-IT, Access Type=Open-Weight Models2026.0379
STARTarget Model=Claude 3.5 Sonnet, Access Type=Closed-Source Models2026.0374
X-teamingTarget Model=Claude 3.5 Sonnet, Access Type=Closed-Source Models2026.0367.9
ActorAttackTarget Model=Claude 3.5 Sonnet, Access Type=Closed-Source Models2026.0366.5
CrescendoTarget Model=Llama 3-70B-IT, Access Type=Open-Weight Models2026.0362
CrescendoTarget Model=Llama 3-8B-IT, Access Type=Open-Weight Models2026.0360
CrescendoTarget Model=Claude 3.5 Sonnet, Access Type=Closed-Source Models2026.0350
CrescendoTarget Model=GPT 4o, Access Type=Closed-Source Models2026.0346
ActorAttackTarget Model=Gemini 2.0-Flash, Access Type=Closed-Source Models2026.0342.1
CoATarget Model=Llama 3-8B-IT, Access Type=Open-Weight Models2026.0325.5
CoATarget Model=Llama 3-70B-IT, Access Type=Open-Weight Models2026.0318.8
CoATarget Model=GPT 4o, Access Type=Closed-Source Models2026.0317.5
CoATarget Model=Claude 3.5 Sonnet, Access Type=Closed-Source Models2026.033.4