OpenAI releases GPT-6 Cyber with critical cybersecurity capability rating

OpenAI released GPT-6 Cyber on November 6, 2026, marking the first artificial intelligence model the company has rated as critical level for cyber capability under its Preparedness Framework. The designation represents OpenAI's highest internal security classification for AI-assisted cybersecurity threats.

The release occurs amid intensifying debate over AI security risks and regulatory frameworks following the Democratic midterm election victory. A critical cyber capability rating indicates the model could significantly accelerate offensive cybersecurity operations, raising questions about access controls, dual-use technology governance, and the adequacy of existing AI safety protocols as models approach or exceed human expert performance in security-sensitive domains.

Sources

AI Daily Digest, GitHub issue 205, published November 7, 2026, github.com