Prompt classification on OpenAIM
81.1F1Qwen3Guard-Gen
Evaluation results
| Method | Links | |
|---|---|---|
| Qwen3Guard-GenModel Size=8B, Evaluation protocol=loose2026.01 | 81.1 | |
| Qwen3Guard-GenModel Size=4B, Evaluation protocol=loose2026.01 | 80.7 | |
| GPT-OSS-SafeGuardModel Size=20B2026.01 | 80.5 | |
| ShieldGemmaModel Size=9B2026.01 | 79.6 | |
| Llama3GuardModel Size=8B2026.01 | 79 | |
| Qwen3Guard-GenModel Size=0.6B, Evaluation protocol=loose2026.01 | 77.5 | |
| NemotronGuardV2Model Size=8B2026.01 | 75.7 | |
| NemotronReasoningModel Size=4B2026.01 | 75.6 | |
| PolyGuardModel Size=7B2026.01 | 75.5 | |
| Llama4GuardModel Size=12B2026.01 | 73.6 | |
| YuFeng-XGuardModel Size=0.6B2026.01 | 72.7 | |
| WildGuardModel Size=7B2026.01 | 72.3 | |
| YuFeng-XGuardModel Size=8B2026.01 | 72.3 | |
| Qwen3Guard-GenModel Size=4B, Evaluation protocol=strict2026.01 | 68.1 | |
| Qwen3Guard-GenModel Size=8B, Evaluation protocol=strict2026.01 | 68.1 | |
| Qwen3Guard-GenModel Size=0.6B, Evaluation protocol=strict2026.01 | 65.8 |