OpenAI has launched a new cybersecurity initiative called Daybreak, using its large language models, Codex’s agentic capabilities and security partnerships to identify software vulnerabilities and help organizations respond to cyber risks. The initiative brings OpenAI into closer competition with other AI providers developing security tools for enterprise customers.
AI Models Move Deeper Into Cyber Defense
Daybreak is designed to use AI reasoning to identify and prioritize high-impact threats, with the goal of helping companies detect weak points in software and act on them more quickly. OpenAI said the initiative will work with industry and government partners to build autonomous cyber defense capabilities into software from the beginning.
Registration Begins for FutureCrime Summit 2026, India’s Largest Cybercrime Conference
The launch comes after Anthropic introduced Mythos in limited preview last month. The model drew attention after highlighting major vulnerabilities across software used in several industries. Anthropic had earlier sought technology vendors to support its Project Glasswing effort around Mythos.
Cloudflare, Cisco, CrowdStrike Join Early Users
OpenAI said Cloudflare, Cisco, CrowdStrike, Oracle and Zscaler are among the companies already using Daybreak. Unlike Anthropic’s Mythos, Daybreak is publicly available, and companies can request an assessment of their security risks.
The system is meant to operate in three stages. It first prioritizes threats using AI reasoning and token usage, then generates and tests risks within an enterprise environment under scoped access, monitoring and review. In the final stage, it sends audit-ready evidence to help companies track, validate and fix vulnerabilities.
Analysts See Complementary Role for AI Security Tools
Industry analysts said AI could help reduce delays between finding vulnerabilities, developing fixes and deploying them, especially as adversaries also use AI to scale attacks. Jeff Pollard, vice president and principal analyst at Forrester, said technology and cybersecurity innovation leaders should test these capabilities to understand what they offer.
John Watts, vice president analyst at Gartner, said Daybreak competes more directly with application security, posture management and AI-enabled application security testing tools, but is more likely to complement such systems than fully replace them. He said organizations still need resources across the remediation chain, including patch testing, deployment and rollback, to reduce operational impact.
The launch also comes as OpenAI expands its enterprise push. The company launched a standalone consulting business on Monday to help organizations adopt AI, deploy forward-engineering teams and strengthen its AI talent base through acquisitions. Anthropic also recently launched its own enterprise AI services company with a group of private equity companies.
