Jailbreak evaluation on abuse, disinformation, hate prompts
99.9Not Unsafe Rategpt-5-thinking
Evaluation results
| Method | Links | |
|---|---|---|
| gpt-5-thinking2025.12 | 99.9 | |
| OpenAI o32025.12 | 99.5 | |
| GPT-4o2025.12 | 98.1 | |
| gpt-5-main2025.12 | 97.8 |
| Method | Links | |
|---|---|---|
| gpt-5-thinking2025.12 | 99.9 | |
| OpenAI o32025.12 | 99.5 | |
| GPT-4o2025.12 | 98.1 | |
| gpt-5-main2025.12 | 97.8 |