OpenAI has just announced the launch of Daybreak, a cybersecurity initiative that looks to be the company’s competitor to Anthropic’s Project Glasswing.
To add some context, Project Glasswing uses Anthropic’s unreleased AI model, Claude Mythos Preview, to meet its clients’ cybersecurity needs. According to Engadget, Mozilla revealed that Mythos helped it find and patch 271 vulnerabilities within the latest release of its Firefox browser.
In the announcement, OpenAI explained that, beyond using its various AI models, Daybreak is designed around cybersecurity, which should be built into the software from the start rather than centring on finding and fixing vulnerabilities. It’s also noted that Daybreak aims to focus on high-impact issues first, and reduce the time it takes to generate and test patches within repositories and to send back results with audit-ready evidence to clients from hours to mere minutes.
Introducing Daybreak: frontier AI for cyber defenders.
Daybreak brings together the most capable OpenAI models, Codex, and our security partners to accelerate cyber defense and continuously secure software.
A step toward a future where security teams can move at the speed… pic.twitter.com/AGfXhmJb5E
— OpenAI (@OpenAI) May 11, 2026
Daybreak will use GPT-5.5 for general purposes and GPT-5.5 with Trusted Access for Cyber and most defensive security workflows, including “secure code review, vulnerability triage, malware analysis, detection engineering and patch validation.” Daybreak will also rely on GPT-5.5-Cyber for other methods, such as preview testing for specialized workflows, controlled validation, and penetration testing.
OpenAI is already working with several partners, including Cloudflare, Cisco, Cloudstrike, Palo Alto Networks, Oracle, and Akamai.
