Share your thoughts, 1 month free Claude Pro on usSee more

Prompt Injection and Jailbreak Detection on WARD (test)

99.33Accuracy

WARD-0.8b

Updated 19d ago

Evaluation Results

Method	Links
WARD-0.8b 2026.05		99.33	99.6	99.07	99.34
WARD-2b 2026.05		99.2	99.67	98.75	99.2
Gemini-3-Flash 2026.05		96.37	96.53	96.21	96.37
Gemma-4-31B 2026.05		95.5	91.67	99.28	95.32
GPT-5.4-Mini 2026.05		95.27	96.6	94.09	95.33
Gemini-3.1-pro 2026.05		94.7	96.2	93.4	94.78
GPT-5.4 2026.05		93.47	92.07	94.72	93.37
WebAgentGuard-8b 2026.05		91.97	85.2	98.54	91.38
Claude-Haiku-4.5 2026.05		90.86	98.55	84.49	90.98
Claude-Sonnet-4.6 2026.05		90.43	97.4	85.49	91.06
WebAgentGuard-4b 2026.05		88.3	78.4	97.76	87.01
Qwen-3.5-4B 2026.05		87.37	86.47	88.05	87.25
Qwen3.5-35B-A3B 2026.05		86.03	85.53	86.4	85.96
Qwen-3.5-2B 2026.05		82.5	85.2	80.83	82.96
Gemma-3-4B 2026.05		78.67	84.67	75.6	79.87
GPT-Oss-Safeguard-20b 2026.05		70.33	40.67	100	57.82
Qwen-3.5-0.8B 2026.05		68.4	73.67	66.65	69.98
SuperAgent-Guard-4b 2026.05		68.4	37.07	99.29	53.98
Llama-Guard-4-12B 2026.05		61.27	23.2	97.21	37.46
BrowseSafe 2026.05		59.7	19.4	100	32.5
Prompt-Guard-1-86M 2026.05		57.43	32.13	65.05	43.02
SuperAgent-Guard-1.7b 2026.05		56.57	17.2	80.88	28.37
DataSentinel 2026.05		56.5	42.6	59	49.48
Prompt-Guard-2-86M 2026.05		54.7	9.4	100	17.18
PromptArmor 2026.05		54.33	53.33	54.42	53.87
Llama-Guard-3-Vision-11B 2026.05		50	0	0	0
GuardReasoner-VL-7B 2026.05		50	0	0	0