OpenAI releases GPT-6 Cyber as its first critical-rated cybersecurity model
On Thursday, November 5, 2026, OpenAI launched GPT-6 Cyber, a model designed to discover and exploit security vulnerabilities, making it available to a limited group of vetted enterprise clients. It is the company’s first model in history to be classified in the critical risk category after successfully discovering 81 vulnerabilities in the Firefox browser and building exploits for 948 vulnerabilities in the Linux kernel.