Agent Evaluation on Auto-ClawEval
93.3SafetyGPT-5-nano
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| GPT-5-nanoFamily=OpenAI2026.04 | 93.3 | 48.9 | 100 | 54.9 | |
| MiniMax M2.5Family=MiniMax2026.04 | 93 | 35.5 | 100 | 43.6 | |
| GPT-5.4Family=OpenAI2026.04 | 91 | 56.7 | 100 | 58.8 | |
| MiniMax M2.7Family=MiniMax2026.04 | 90.5 | 43.8 | 100 | 49.4 | |
| Claude Sonnet 4.6Family=Anthropic2026.04 | 90.3 | 50 | 100 | 53.7 | |
| GLM 5Family=Zipu AI2026.04 | 90.2 | 45.3 | 100 | 50.1 | |
| GLM 5 TurboFamily=Zipu AI2026.04 | 89 | 46.2 | 100 | 49.8 | |
| Claude Opus 4.6Family=Anthropic2026.04 | 87.3 | 49.7 | 100 | 52.4 |