OpenAI Launches Daybreak to Automate Vulnerability Detection and Patching

OpenAI has unveiled Daybreak, a new initiative designed to help organizations identify and fix security flaws before attackers can exploit them. The system builds on the Codex Security AI agent introduced in March, which analyzes an organization's code to map out potential attack paths, validate likely vulnerabilities, and automatically flag the highest-risk ones.

Daybreak combines GPT-5.5-Cyber - released last week - with Codex Security and other specialized cyber models, including GPT-5.5 with Trusted Access for Cyber. The move comes just over a month after rival Anthropic announced Claude Mythos, a security-focused AI model it deemed too dangerous for public release and shared only privately through its Project Glasswing. Despite those restrictions, unauthorized parties reportedly accessed the model.

Until now, OpenAI lacked a comparable security product. Unlike Glasswing, Daybreak isn't built on a single model. Instead, OpenAI says it integrates its most capable models, Codex, and security partners. The company is also working with industry and government partners as it prepares to deploy increasingly capable cyber models at scale.

Source: The Verge