List your company

Agent Safety Evaluation on ToolEmu

99SafetySafeMCP
30.3648.186683.82Jun 1, 2026

Evaluation results

MethodLinks
SafeMCPAgent=GPT-4o2026.0699226044
SafeMCPAgent=GPT-4o-mini2026.0698155640
RL-GuardAgent=Claude-3.5-Sonnet2026.0694396757
SafeMCPAgent=Claude-3.5-Sonnet2026.0694426859
RL-GuardAgent=GPT-4o2026.068994935
AgentMonitorAgent=Claude-3.5-Sonnet2026.0689285948
RL-GuardAgent=GPT-4o-mini2026.068894935
AgentMonitorAgent=Llama-3.1-8B2026.068834531
SafeMCPAgent=Llama-3.1-8B2026.068574633
AgentMonitorAgent=GPT-4o-mini2026.0683104735
AgentMonitorAgent=GPT-4o2026.0681225243
ChainGuardAgent=Claude-3.5-Sonnet2026.0677405955
RL-GuardAgent=Llama-3.1-8B2026.067734030
NeMoGuard-8BAgent=Claude-3.5-Sonnet2026.0676436056
Qwen3Guard-8BAgent=Claude-3.5-Sonnet2026.0672435755
ChainGuardAgent=GPT-4o2026.0671154336
Llama Guard 3Agent=Claude-3.5-Sonnet2026.0669475857
ChainGuardAgent=GPT-4o-mini2026.0665113832
SafironAgent=Claude-3.5-Sonnet2026.0665535959
Qwen3Guard-8BAgent=GPT-4o2026.0663214238
w/o defenseAgent=Claude-3.5-Sonnet2026.0663545858
ChainGuardAgent=Llama-3.1-8B2026.066053327
NeMoGuard-8BAgent=GPT-4o2026.0658224037
Qwen3Guard-8BAgent=GPT-4o-mini2026.0658153633
NeMoGuard-8BAgent=Llama-3.1-8B2026.065273026
Llama Guard 3Agent=Llama-3.1-8B2026.065172926
Qwen3Guard-8BAgent=Llama-3.1-8B2026.064962825
SafironAgent=GPT-4o2026.0645243433
SafironAgent=GPT-4o-mini2026.0645173129
Llama Guard 3Agent=GPT-4o-mini2026.0644173129
NeMoGuard-8BAgent=GPT-4o-mini2026.0643142927
w/o defenseAgent=GPT-4o2026.0642253433
w/o defenseAgent=GPT-4o-mini2026.0642173028
Llama Guard 3Agent=GPT-4o2026.0641203130
SafironAgent=Llama-3.1-8B2026.063672220
w/o defenseAgent=Llama-3.1-8B2026.063372019