OpenAI has released GPT-6 Cyber, a security-focused frontier model designed to support applied security research and defensive engineering. The model, announced November 5, builds on the GPT-6 family but is tuned specifically for threat analysis, vulnerability research, and secure development workflows rather than general-purpose chat tasks.
OpenAI said the model has been used to assist research on the Linux kernel and is already deployed in early defense programs at organizations including Mozilla, Palo Alto Networks, and Cisco. Access is initially limited to select enterprise users through the company's API and Codex environment, with a broader rollout planned in the coming months.
The release reflects a widening pattern across the AI industry as labs ship purpose-built systems for vertical markets. Analysts said a specialized security model could shift how enterprises evaluate AI procurement, moving the conversation from generic assistants toward tools trained for a specific regulatory and operational domain.
Image source: i.ibb.co