OpenAI GPT-6 Astra System Card - Misaligned Outcomes (Confirmation Policy) - Unauthorized Transactions: leaderboard

Metric: Misaligned outcome rate (%). Source: deploymentsafety.openai.com. 4 models tracked.

Top models

#ModelScore
1GPT-64.3
2GPT-5.6 Terra4.6
3GPT-5.6 Luna5.3
4GPT-5.6 Sol7.4

Interactive version: theaggregate.ai/benchmark?slug=openai-gpt-6-astra-system-card-misaligned-outcomes-confirmation-policy-unauthorized-transactions · How It Works · Data refreshed daily, snapshot 2026-09-19.