Input Moderation on AEGIS (test)
91.4F1 ScoreQwen3Guard-8B-Gen-strict
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Qwen3Guard-8B-Gen-strictStreaming=no2026.04 | 91.4 | — | — | — | — | — | |
| Qwen3Guard-0.6B-Gen-strictStreaming=no2026.04 | 90.8 | — | — | — | — | — | |
| Qwen3Guard-4B-Gen-strictStreaming=no2026.04 | 90.8 | — | — | — | — | — | |
| PolyGuard-Qwen-7BStreaming=no2026.04 | 90.3 | — | — | — | — | — | |
| WildGuard-7BStreaming=no2026.04 | 89.4 | — | — | — | — | — | |
| Llama3-StreamGuard-1BStreaming=yes2026.04 | 89 | — | — | — | — | — | |
| Llama3-StreamGuard-8BStreaming=yes2026.04 | 88.5 | — | — | — | — | — | |
| Llama3-StreamGuard-3BStreaming=yes2026.04 | 87.1 | — | — | — | — | — | |
| Qwen3Guard-Stream-4B-strictStreaming=yes2026.04 | 85.9 | — | — | — | — | — | |
| Qwen3Guard-Stream-8B-strictStreaming=yes2026.04 | 85.7 | — | — | — | — | — | |
| Qwen3Guard-Stream-0.6B-strictStreaming=yes2026.04 | 85.2 | — | — | — | — | — | |
| NemoGuard-8BStreaming=no2026.04 | 81.4 | — | — | — | — | — | |
| Qwen3Guard-Stream-0.6B-looseStreaming=yes2026.04 | 77.7 | — | — | — | — | — | |
| Qwen3Guard-0.6B-Gen-looseStreaming=no2026.04 | 76.9 | — | — | — | — | — | |
| Qwen3Guard-4B-Gen-looseStreaming=no2026.04 | 76.3 | — | — | — | — | — | |
| Qwen3Guard-8B-Gen-looseStreaming=no2026.04 | 76 | — | — | — | — | — | |
| Qwen3Guard-Stream-4B-looseStreaming=yes2026.04 | 75.5 | — | — | — | — | — | |
| Qwen3Guard-Stream-8B-looseStreaming=yes2026.04 | 75.5 | — | — | — | — | — | |
| LlamaGuard3-8BStreaming=no2026.04 | 71.5 | — | — | — | — | — | |
| ShieldGemma-9BStreaming=no2026.04 | 70.3 | — | — | — | — | — | |
| ShieldGemma-27BStreaming=no2026.04 | 69 | — | — | — | — | — | |
| CHAIRO2026.04 | 68 | 81.5 | 81.6 | 82.5 | 40.7 | 53.9 | |
| LlamaGuard4-12BStreaming=no2026.04 | 67.8 | — | — | — | — | — | |
| SFT2026.04 | 61.6 | 74.7 | 71.4 | 80 | 26.7 | 55 | |
| RAG2026.04 | 54.7 | 62.3 | 72.3 | 68.3 | 27.6 | 42.9 | |
| Qwen3-8B2026.04 | 52.9 | 67.2 | 65.2 | 66.4 | 25.8 | 40 |