List your company

Agentic Safety Moderation on AgentHazard

109ASRw/o guardrail
68.4478.9789.5100.03May 28, 2026

Evaluation results

MethodLinks
w/o guardrailGuardrail=None2026.05109——————
Qwen3Guard-Gen-4BGuardrail=Qwen3Guard-Gen-4B2026.0510900.12080.32260.01270.012913.02
Llama-Guard-3-8BGuardrail=Llama-Guard-3-8B2026.0510900.10080.32640.01180.01243
AgentDoG 1.5-0.8BBase Model=Qwen3.5 Thinking Base, Size=0.8B2026.0576-12.690.18920.46030.01480.0149358.44
AgentDoG 1.5-4BBase Model=Qwen3.5 Thinking Base, Size=4B2026.0570-150.2960.73780.02070.0209493.38