OpenAI introduces GPT-6 Cyber for approved Daybreak Red customers

OpenAI's post dated 5 November 2026 introduces GPT-6 Cyber. It says the model is for finding and validating exploitable vulnerabilities, and for staying within the scope a researcher sets. OpenAI classifies it as Critical in cybersecurity under its Preparedness Framework, up from High for GPT-5.6 Cyber. The post does not state a token price or a context window.

On ExploitGym, the post says GPT-6 Cyber scores 48.5%, up from 32.1% for GPT-5.6 Cyber, which it also describes as a 51% relative improvement and about 16 percentage points. On a separate alignment test, it says the model avoided exploiting out-of-scope services in 99.93% of tests without its deployment prompt, up from 94.49% for GPT-5.6 Cyber, and in 100% of tests with that prompt. On a web test without source code, it says recall was 40.3%, compared with 32.0% for GPT-6 Astra. On RemoteExploitBench, it says GPT-6 Astra running with Daybreak Blue refused 100% of tasks, while GPT-6 Cyber completed 95%.

The post says OpenAI's researchers used the model to develop 948 privilege escalation exploits in Linux, after reproducing 1,724 vulnerabilities. It says that is more than five times the 166 exploits developed by GPT-5.6 Sol with the same research harness, and that 304 critical findings were disclosed to maintainers with candidate patches. It says the model identified 81 previously unknown high-severity vulnerabilities in Firefox, and that an additional 30 known issues were more severe than previously reported. Those counts are OpenAI's.

The post says GPT-6 Cyber is initially available through the API and Codex to a select group of approved Daybreak Red customers, with access to be expanded to all Daybreak Red customers. It also says GPT-5.6 Cyber is being made available to all Daybreak Red customers. The API changelog for 5 November 2026 lists the model as gpt-6-cyber and says to see Daybreak setup before the first request. The post describes a new group, Daybreak Red for Critical Infrastructure, for approved defenders of critical systems. It refers to the existing $1 billion Daybreak for Frontline Defenders program. That program was already described in OpenAI's 2 November security post.

What it means for business: this is not a general ChatGPT release. The post limits first access to approved Daybreak Red customers, and it does not state a price. OpenAI's own rating is Critical for cybersecurity. This describes the posts. It is not advice on connecting the model to a system.