List your company

Prompt Injection and Jailbreak Detection on WARD (test)

99.33AccuracyWARD-0.8b
48.026861.345974.66587.9841May 14, 2026

Evaluation results

MethodLinks
WARD-0.8bModel Category=Ours, Number of Parameters=0.8b2026.0599.3399.699.0799.34
WARD-2bModel Category=Ours, Number of Parameters=2b2026.0599.299.6798.7599.2
Gemini-3-FlashModel Category=Closed-source API2026.0596.3796.5396.2196.37
Gemma-4-31BModel Category=Open-source instructed model, Number of Parameters=31B2026.0595.591.6799.2895.32
GPT-5.4-MiniModel Category=Closed-source API2026.0595.2796.694.0995.33
Gemini-3.1-proModel Category=Closed-source API2026.0594.796.293.494.78
GPT-5.4Model Category=Closed-source API2026.0593.4792.0794.7293.37
WebAgentGuard-8bModel Category=Guard model (Prompt Injection), Number of Parameters=8b2026.0591.9785.298.5491.38
Claude-Haiku-4.5Model Category=Closed-source API2026.0590.8698.5584.4990.98
Claude-Sonnet-4.6Model Category=Closed-source API2026.0590.4397.485.4991.06
WebAgentGuard-4bModel Category=Guard model (Prompt Injection), Number of Parameters=4b2026.0588.378.497.7687.01
Qwen-3.5-4BModel Category=Open-source instructed model, Number of Parameters=4B2026.0587.3786.4788.0587.25
Qwen3.5-35B-A3BModel Category=Open-source instructed model, Number of Parameters=35B2026.0586.0385.5386.485.96
Qwen-3.5-2BModel Category=Open-source instructed model, Number of Parameters=2B2026.0582.585.280.8382.96
Gemma-3-4BModel Category=Open-source instructed model, Number of Parameters=4B2026.0578.6784.6775.679.87
GPT-Oss-Safeguard-20bModel Category=Guard model (Jail Break), Number of Parameters=20b2026.0570.3340.6710057.82
Qwen-3.5-0.8BModel Category=Open-source instructed model, Number of Parameters=0.8B2026.0568.473.6766.6569.98
SuperAgent-Guard-4bModel Category=Guard model (Prompt Injection), Number of Parameters=4b2026.0568.437.0799.2953.98
Llama-Guard-4-12BModel Category=Guard model (Jail Break), Number of Parameters=12B2026.0561.2723.297.2137.46
BrowseSafeModel Category=Guard model (Prompt Injection)2026.0559.719.410032.5
Prompt-Guard-1-86MModel Category=Guard model (Jail Break), Number of Parameters=86M2026.0557.4332.1365.0543.02
SuperAgent-Guard-1.7bModel Category=Guard model (Prompt Injection), Number of Parameters=1.7b2026.0556.5717.280.8828.37
DataSentinelModel Category=Guard model (Prompt Injection)2026.0556.542.65949.48
Prompt-Guard-2-86MModel Category=Guard model (Jail Break), Number of Parameters=86M2026.0554.79.410017.18
PromptArmorModel Category=Guard model (Prompt Injection)2026.0554.3353.3354.4253.87
Llama-Guard-3-Vision-11BModel Category=Guard model (Jail Break), Number of Parameters=11B2026.0550000
GuardReasoner-VL-7BModel Category=Guard model (Jail Break), Number of Parameters=7B2026.0550000