OpenAI introduces GPT-6 Cyber for security research, for select Daybreak Red customers

OpenAI introduces GPT-6 Cyber, a security-research model that it says pairs stronger capabilities with better alignment. It says this is the model behind the Linux kernel research it announced the day before, and that security teams at Mozilla, Palo Alto Networks, Calif and Cisco, as well as OpenAI's own, are already using it.

OpenAI reports that GPT-6 Cyber scores 48.5% on ExploitGym, up from 32.1% for GPT-5.6 Cyber, and 40.3% recall on a new web penetration-testing evaluation, against 32.0% for GPT-6 Astra. On a new benchmark of known vulnerabilities in applications such as GitLab, PostgreSQL and Jenkins, it says GPT-6 Cyber completed 95% of tasks while GPT-6 Astra under Daybreak Blue refused all of them. On an evaluation of staying within scope, it says the model avoided exploiting out-of-scope services in 99.93% of tests without deployment prompts (94.49% for GPT-5.6 Cyber) and 100% with them. OpenAI classifies it as Critical in cybersecurity under its Preparedness Framework, up from High for GPT-5.6 Cyber, and has published a system card.

OpenAI says its researchers used the model to develop working exploits for 948 Linux kernel vulnerabilities and to find 81 previously unknown high-severity Firefox vulnerabilities, for which Mozilla has begun releasing fixes. Customers it quotes include Mozilla, Calif, Cognition, Palo Alto Networks, Cisco, SentinelOne, Perplexity and CodeRabbit.

GPT-6 Cyber is available to select Daybreak Red customers, and OpenAI says it will expand access over time. OpenAI's API pricing page lists gpt-6-cyber at 20 input, 100 output and 2 cached input, and its API changelog tells customers to see the Daybreak setup before sending a first request. It describes Daybreak Red as enabling offensive security with deeper exploit development and authorized testing, and Daybreak Blue as helping verified defenders validate vulnerabilities and fixes with fewer refusals. Safeguards include IP allowlisting for Daybreak Red API access by default, and Private Safety Processing for API customers using Zero Data Retention. OpenAI adds a Daybreak Red for Critical Infrastructure group for approved defenders, expands the Daybreak Blue alpha for GPT-6 Astra to more organizations and individual researchers, and says it wants to extend Daybreak to organizations of every size, independent researchers and open-source maintainers.

OpenAI has scheduled a "Cyber at the Frontier" livestream on November 12 at 10am PT for more cyber announcements. The capability and alignment claims are OpenAI's own.