OpenAI GPT-6 Astra System Card - Image Input - Self-Harm: leaderboard
Table 4 (image input evaluations) of OpenAI's GPT-6 Astra System Card (Deployment Safety Hub, 2026-09-03): safe completion rate on disallowed self-harm requests that combine text and images, the suite introduced with ChatGPT Agent. Percent (the card prints 0-1), higher is better. Scores sit near the ceiling. GPT-6 Sol and GPT-6 Luna come from Table 26 of the card's section 11 (GPT-6 Sol and GPT-6 Luna, added 2026-09-22), which re-prints the other cells.
Metric: Safe completion rate (%). Source: deploymentsafety.openai.com. Source: deploymentsafety.openai.com. Source: deploymentsafety.openai.com. Saturation forecast: Estimated already saturated. 2 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-6 Luna | 99.9 | #61 |
| 2 | GPT-6 Astra | 99.7 | #6 |
| 3 | GPT-5.4 (Thinking) | 99.6 | #83 (GPT-5.4) |
| 4 | GPT-6.1 Sol | 99.4 | #7 |
| 5 | GPT-5.6 Sol | 98.9 | #22 |
| 6 | GPT-6.1 Terra | 98.7 | #15 |
| 7 | GPT-5.5 | 98.3 | #33 |
| 8 | GPT-6 Sol | 98.2 | #19 |
No result here: #3 Claude Opus 5.5, #9 Gemini 4 Argon, #10 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=openai-gpt-6-astra-system-card-image-input-self-harm · How It Works · Data refreshed daily, snapshot 2026-11-05.