OpenAI has launched GPT-6 Cyber, a model built for cybersecurity work, according to Reuters. The report states the release carries a "Critical" cybersecurity rating. The launch was announced on Thursday, November 5, 2026.
OpenAI described the model as an advance in aligned cybersecurity agents, according to its own announcement. Reporting from Axios says the model identified ways to attack close to 1,000 Linux vulnerabilities. The Hacker News reported that an OpenAI pipeline generated 948 Linux root escalation exploits during a 30-hour run. Those figures come from the outlets' coverage and have not been independently verified by CYBEV.io.
The "Critical" rating is notable because it suggests the model's offensive security capabilities were judged significant enough to flag. Reuters did not provide further detail in the summary available. The release also arrives alongside separate security reporting: The Information published an exclusive about a researcher finding a vulnerability in software used for data centers. There is no indication in the material that the two stories are connected.
What to watch next: whether OpenAI publishes fuller details on the "Critical" rating and how access to GPT-6 Cyber will be restricted.
