OpenAI announced GPT-6 Cyber on 5 November 2026, describing it as a “major leap in security research.” The specialised AI model is designed to help cybersecurity professionals identify vulnerabilities, develop exploits and strengthen software defences, combining advanced vulnerability detection with improved safeguards to keep it within authorised boundaries. OpenAI says the model has already helped uncover thousands of vulnerabilities in widely used software, including Linux and Firefox; researchers used it to develop 948 privilege-escalation exploits in Linux and identify 81 previously unknown high-severity vulnerabilities in Firefox, and the research also helped produce candidate patches, with Mozilla subsequently releasing fixes for many of the identified issues. GPT-6 Cyber scored 48.5% on the ExploitGym benchmark, compared with 32.1% for its predecessor, GPT-5.6 Cyber, a 51% relative improvement. Johannes Landgraf, Head of Cyber at OpenAI, called it “a major step forward for security research and defenders” and said the company invested heavily in making the model trustworthy, training it to stay within the scope set by its operator. In alignment evaluations, the model avoided exploiting out-of-scope services in 99.93% of tests without deployment safety prompts, rising to 100% with these prompts enabled, the company says.