OpenAI introduced GPT-6 Cyber on November 5 for a selected group of approved Daybreak Red customers through its API and Codex. Daybreak Red is OpenAI’s access tier for advanced, authorized security testing. Wider access for Daybreak Red customers is planned, rather than available to every subscriber at launch.
GPT-6 Cyber is a security-specialized version of GPT-6 Astra, trained for authorized vulnerability discovery, validation and testing. OpenAI rates it Critical in cybersecurity, the same designation as Astra and up from High for GPT-5.6 Cyber.
Capability Gains Have Evaluation Boundaries
OpenAI reports an ExploitGym score of 48.5%, compared with 32.1% for GPT-5.6 Cyber. The evaluation measures exploiting specified vulnerabilities in controlled targets; it does not establish success rates across arbitrary production systems.
The company also reports better compliance with authorized scope in its alignment tests. Those results do not remove the need for sandboxing, action review and human oversight.
Approval, Pricing and Data Handling
Daybreak Red approval alone does not guarantee GPT-6 Cyber access. Organizations need separate model approval and security controls including identity verification, single sign-on, multi-factor authentication and API IP allowlisting.
The model documentation lists a 1,050,000-token context window and up to 128,000 output tokens. Standard short-context API rates per million tokens are $20 for input, $2 for cached input and $100 for output. Cache writes cost $25 for 30 minutes or $50 for 12 hours; tool use can carry separate charges.
Standard retention remains the default. API customers using Zero Data Retention generally need Private Safety Processing, which checks encrypted interaction records kept in customer storage for 30 days. Daybreak is for work on systems the user owns or is explicitly authorized to test; it does not permit reselling or proxying access to third-party customers.