OpenAI GPT-6 Astra System Card - Challenging Prompts - Self-Harm: leaderboard
Metric: Safe completion rate (%). Source: deploymentsafety.openai.com. 4 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-6 | 99.2 |
| 2 | GPT-5.4 (Thinking) | 95.9 |
| 3 | GPT-5.6 Sol | 94.5 |
| 4 | GPT-5.5 | 91.7 |
Interactive version: theaggregate.ai/benchmark?slug=openai-gpt-6-astra-system-card-challenging-prompts-self-harm · How It Works · Data refreshed daily, snapshot 2026-09-19.