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

Safety Detection on AgentHazard Strongest

90.87Accuracy

BraveGuard-Qwen3-Guard-4B

Updated 1d ago

Evaluation Results

Method
BraveGuard-Qwen3-Guard-4B 2026.05	90.87	92.21	94.94
BraveGuard-Qwen3-Guard-8B 2026.05	90.49	91.8	94.71
BraveGuard-Qwen3-Guard-4B 2026.05	90.49	91.19	94.31
BraveGuard-Llama-Guard-8B 2026.05	89.35	90.75	93.64
BraveGuard-Llama-Guard-8B 2026.05	87.83	90.57	93.25
BraveGuard-Qwen3-Guard-8B 2026.05	87.45	87.67	92.34
GPT-5.4 2026.05	84.41	88.11	91.3
BraveGuard-Qwen3-Guard-8B 2026.05	83.65	91.28	89.22
BraveGuard-Llama-Guard-8B 2026.05	82.89	82.89	87.32
BraveGuard-Llama-Guard-8B 2026.05	82.51	92.82	88.73
BraveGuard-Qwen3-Guard-8B 2026.05	81.37	80.21	85.96
AgentDoG-Llama3.1-8B 2026.05	80.99	82.38	88.94
BraveGuard-Qwen3-Guard-4B 2026.05	80.99	88.72	87.37
GPT-5.4 2026.05	80.61	85.9	88.44
BraveGuard-Qwen3-Guard-4B 2026.05	80.61	82.35	85.79
AgentDoG-Qwen2.5-7B 2026.05	79.09	80.74	87.75
Gemini 3.1 Flash 2026.05	77.95	79.51	87
Claude Sonnet 4.6 2026.05	76.43	77.87	85.97
GPT-5.4 2026.05	74.14	85.13	83
GPT-5.4 2026.05	73.38	82.89	81.58
Gemini 3.1 Flash 2026.05	71.86	71.37	81.41
Claude Sonnet 4.6 2026.05	69.2	66.08	78.74
AgentDoG-Llama3.1-8B 2026.05	66.92	63.88	76.92
AgentDoG-Llama3.1-8B 2026.05	66.16	58.82	71.2
AgentDoG-Qwen2.5-7B 2026.05	65.02	60.51	71.95
AgentDoG-Qwen2.5-7B 2026.05	65.02	60.79	75
AgentDoG-Llama3.1-8B 2026.05	64.26	58.97	70.99
Claude Sonnet 4.6 2026.05	63.12	56.15	68.4
AgentDoG-Qwen2.5-7B 2026.05	61.98	54.55	67.11
Gemini 3.1 Flash 2026.05	59.7	50.8	64.19
Claude Sonnet 4.6 2026.05	57.79	53.33	65.2
Gemini 3.1 Flash 2026.05	57.41	53.33	65
Qwen3.5 Flash 2026.05	50.95	47.13	64.07
Qwen3.5 Flash 2026.05	40.68	18.18	30.36
YuFeng-XGuard 2026.05	38.78	17.11	28.44
Qwen3.5 Flash 2026.05	36.88	26.87	42.36
Qwen3.5 Flash 2026.05	33.59	13.92	23.68
YuFeng-XGuard 2026.05	33.46	13.33	22.91
Qwen3-Guard-4B 2026.05	30.04	1.6	3.16
Qwen3-Guard-8B 2026.05	29.66	1.07	2.12
LlamaGuard3-8B 2026.05	29.28	0.53	1.06
NemoGuard 2026.05	28.94	2.73	4.21
NemoGuard 2026.05	28.52	11.76	18.97
YuFeng-XGuard 2026.05	27.38	21.72	35.69
YuFeng-XGuard 2026.05	27.38	16.3	27.92
LlamaGuard3-8B 2026.05	27	3.59	6.8
Qwen3-Guard-4B 2026.05	26.62	1.03	2.03
Qwen3-Guard-8B 2026.05	26.24	1.03	2.02
LlamaGuard3-8B 2026.05	16.35	3.08	5.98
Qwen3-Guard-8B 2026.05	15.76	2.97	5.54
Qwen3-Guard-4B 2026.05	15.59	2.2	4.31
NemoGuard 2026.05	15.21	3.96	7.47
LlamaGuard3-8B 2026.05	13.69	6.97	13.03
NemoGuard 2026.05	12.17	5.74	10.81
Qwen3-Guard-8B 2026.05	11.03	4.1	7.87
Qwen3-Guard-4B 2026.05	8.37	1.23	2.43