Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Text Query Safety on SorryBench
Loading...
89.43
F1 Score
GraniteGuardian
45.1468
56.6434
68.14
79.6366
Jun 22, 2026
F1 Score
Updated 1mo ago
Evaluation Results
Method
Method
Links
F1 Score
GraniteGuardian
2026.06
89.43
GPT-5.1
2026.06
82.87
YuFeng-XGuard-Reason-8B
2026.06
79.99
Qwen3Guard-8B
Policy=strict
2026.06
79.57
SingGuard-8B
Parameters=8B
2026.06
79.12
Gemini3-Pro
2026.06
78.75
SingGuard-2B
Parameters=2B
2026.06
78.5
SingGuard-4B
Parameters=4B
2026.06
77.69
Qwen3-VL-235B
2026.06
76.17
GuardReasoner-VL-7B
2026.06
67.13
Qwen3Guard-8B
Policy=loose
2026.06
61.89
Llama Guard 3
2026.06
59.38
WildGuard
2026.06
58.98
ShieldGemma-9B
2026.06
46.85
Feedback
Search any
task
Search any
task