OpenAI recently launched Daybreak, a cybersecurity initiative that is apparently the company’s competitor to Anthropic’s Project Glasswing. If you recall, Glasswing uses Anthropic’s unreleased AI model, Cloud Mythos Preview, to provide its customers’ cyber defense needs. It’s been promising so far: Mozilla revealed in April that Mythos helped find and fix 271 vulnerabilities in the latest release of the Firefox browser. OpenAI says Daybreak uses a variety of its AI models, including its specialized security agent Codex.
In its announcement, the company explained that Daybreak is built on the premise that cybersecurity should be built into software from the beginning, not just revolve around finding and fixing vulnerabilities. Daybreak aims to prioritize high-impact issues and reduce hours of analysis into minutes, generating and testing patches within the repository and sending results back to customers’ systems with audit-ready evidence. In OpenAI’s example, it asked Codex Security to scan a codebase, verify the highest-risk findings, and fix them.
Find and fix vulnerabilities first with Daybreak pic.twitter.com/yobOSWYeWP
– OpenAI (@OpenAI) 11 May 2026
Daybreak will use GPT-5.5 for general purposes and with trusted access to cyber for most defensive security workflows, including “secure code review, vulnerability triage, malware analysis, detection engineering, and patch verification.” It will also rely on GPT-5.5-Cyber for “preview access to specific workflows, including authorized red teaming, penetration testing, and controlled validation.” OpenAI is already working with several partners on this initiative, including Cloudflare, Cisco, CloudStrike, Palo Alto Networks, Oracle, and Akamai.
<a href