Unsafe Content Categorization on OpenAI Moderation
88.35AccuracyGemini2.5-Flash
Evaluation results
| Method | Links | |
|---|---|---|
| Gemini2.5-Flash2025.12 | 88.35 | |
| GPT4o-mini2025.12 | 84.87 | |
| ProGuard-3B2025.12 | 78.93 | |
| ProGuard-7B2025.12 | 76.63 | |
| LlamaGuard4-12B2025.12 | 69.92 | |
| LlamaGuard2-8B2025.12 | 69.16 | |
| LlamaGuard3-11B-Vision2025.12 | 69.16 | |
| LlamaGuard3-8B2025.12 | 59.96 | |
| LlamaGuard-7B2025.12 | 20.11 |