DropTicks
← Back to AI News OpenAI News • Sep 03, 2026

Safety overview: GPT-6 Astra

GPT-6 Astra is our most capable broadly deployed model and our first to reach the Critical level of cybersecurity capability under our Preparedness Framework.

Login to save Open source Read articles
GPT-6 Astra is our most capable broadly deployed model and our first to reach the Critical level of cybersecurity capability under our Preparedness Framework.

Editorial note: review the original source, verify technical claims, then publish with DropTicks analysis.