List your company
ResearchFormerly WizWand

Safety Classification on Wildguardmix

89.54F1 ScoreSingGuard-4B
-2.541621.364245.2769.1758Dec 23, 2025Jan 22, 2026Feb 21, 2026Mar 23, 2026Apr 22, 2026May 22, 2026Jun 22, 2026
Updated 4mo ago

Evaluation results

MethodLinks
SingGuard-4BParameters=4B2026.0689.54
SingGuard-2BParameters=2B2026.0689.34
Qwen3Guard-8BPolicy=strict2026.0688.95
SingGuard-8BParameters=8B2026.0688.84
WildGuard2026.0688.73
GuardReasoner-VL-7B2026.0688.66
YuFeng-XGuard-Reason-8B2026.0688.07
Qwen3Guard-8BPolicy=loose2026.0685.78
GPT-5.12026.0683.62
Llama Guard 32026.0676.67
Apriel GuardModel Size=8B, Reasoning=false2025.1276
Gemini3-Pro2026.0675.12
Apriel GuardModel Size=8B, Reasoning=true2025.1275
Qwen3-VL-235B2026.0672.55
GraniteGuardian2026.0668.87
ShieldGemma-9B2026.0660.26
IBM Granite GuardianVersion=3.2, Model Size=3B, Reasoning=false2025.1251
IBM Granite GuardianVersion=3.2, Model Size=5B, Reasoning=false2025.1251
IBM Granite GuardianVersion=3.1, Model Size=2B, Reasoning=false2025.1249
Llama Prompt Guard 2Model Size=0.086B, Reasoning=false2025.1248
IBM Granite GuardianVersion=3.3, Model Size=8B, Reasoning=false2025.1246
IBM Granite GuardianVersion=3.3, Model Size=8B, Reasoning=true2025.1239
Llama GuardVersion=3, Reasoning=false2025.1230
Llama GuardVersion=4, Reasoning=false2025.1227
Llama GuardVersion=2, Reasoning=false2025.1226
ShieldGemmaModel Size=9B, Reasoning=false2025.1225
gpt-oss-safeguardModel Size=20B, Reasoning=true2025.1210
Qwen3GuardModel Size=8B, Constraint=loose, Reasoning=false2025.121
Qwen3GuardModel Size=8B, Constraint=strict, Reasoning=false2025.121