The company announced GPT-6 Cyber for approved customers on November 5, a specialized model for cybersecurity research.

On November 5, 2026, OpenAI announced GPT-6 Cyber, a cybersecurity-focused variant of GPT-6 Astra, available through its API and Codex to a select group of approved Daybreak Red customers. OpenAI says it is designed for authorized vulnerability research and exploit development, with safeguards intended to keep work within an assigned scope. The model scored 48.5% on OpenAI’s ExploitGym evaluation, compared with 32.1% for GPT-5.6 Cyber; in a scope test, it avoided exploiting off-limits services in 99.93% of cases. Researchers also used it to find 81 previously unknown high-severity Firefox vulnerabilities and develop working exploits for 948 Linux kernel vulnerabilities, including 588 previously unknown flaws. Those results illustrate both the defensive value and misuse risk of increasingly capable cyber models. OpenAI classifies GPT-6 Cyber at its “Critical” cyber-risk tier and plans to expand access to all Daybreak Red customers; GPT-5.6 Cyber is also being made available to them.