Safety Evaluation on Standard Disallowed Content Evaluation (test)
100Hate Speech Aggregate Accuracygpt-5-thinking
Evaluation results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| gpt-5-thinking2025.12 | 100 | — | 99.1 | 100 | 88.1 | 98.9 | 100 | 100 | 99 | |
| GPT-4o2025.12 | 99.6 | — | 98.3 | 100 | 96.7 | 97.8 | 100 | 100 | 100 | |
| OpenAI o32025.12 | 99.2 | — | 99.1 | 100 | 93 | 92.1 | 100 | 100 | 100 | |
| gpt-5-main2025.12 | 98.7 | — | 99.1 | 99.2 | 98 | 98.9 | 100 | 100 | 100 |