Safety Evaluation on Expert Red Teaming Harmful Conversations - Biological Threats
60Win Rategpt-5-thinking
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| gpt-5-thinkingOpponent=OpenAI o32025.12 | 60 | — | 0.4 | |
| OpenAI o3Opponent=gpt-5-thinking2025.12 | 40 | — | — |
| Method | Links | |||
|---|---|---|---|---|
| gpt-5-thinkingOpponent=OpenAI o32025.12 | 60 | — | 0.4 | |
| OpenAI o3Opponent=gpt-5-thinking2025.12 | 40 | — | — |