List your company

Content Safety on SafetyBench (test)

89.3Macro F1-ScoreGPT-3.5
77.96480.90783.8586.793Jun 26, 2026

Evaluation results

MethodLinks
GPT-3.52026.0689.3
Qwen1.5-72BParameters=72B2026.0687.7
Yuvion-32BParameters=32B2026.0687.5
Qwen2-32BParameters=32B2026.0686.2
DeepSeek-67BParameters=67B2026.0685.9
Qwen2-8BParameters=8B2026.0682.4
Yuvion-8BParameters=8B2026.0681.6
Qwen2-72BParameters=72B2026.0678.4