On 5 November 2026, OpenAI released GPT-6 Cyber as a security research model, with professional access through Daybreak. Capability and scope-adherence results in the announcement are vendor evaluations. Mozilla’s Firefox 159 security advisory, announced on 27 October, provides an independently readable record of some fixes, including reporters attributed to OpenAI.
In a technical write-up dated 9 November 2026, OpenAI described Control Room. Research jobs carry identifiers, and inference engines check permission before producing output. After a job is stopped, related requests are denied. OpenAI says output stops within 35 seconds of an intervention decision; that figure is the company’s own, and this site has not independently measured it.