OpenAI has introduced a new cybersecurity initiative called Daybreak. The platform is designed to help organizations detect software vulnerabilities, validate security patches, and improve protection against cyberattacks using AI.
The company says Daybreak combines its latest AI models with Codex Security to help developers and security teams identify risky code and possible attack paths much earlier in the development process.
According to OpenAI, the platform can perform secure code reviews, threat modeling, dependency risk analysis, vulnerability testing, and remediation guidance. The goal is to make software more secure before attackers get a chance to exploit weaknesses.
Daybreak uses Codex Security to create editable threat models for software repositories. It can analyze codebases, focus on high-risk areas, test vulnerabilities in isolated environments, and even suggest possible fixes for discovered issues.
OpenAI is currently limiting access to the platform. Interested organizations need to request vulnerability scans or contact the company directly for access.
The initiative arrives at a time when AI is rapidly changing the cybersecurity industry. AI-powered tools are now capable of discovering vulnerabilities much faster than traditional research methods. This has created a new challenge where security teams often struggle to patch issues quickly enough.
Earlier this year, HackerOne temporarily paused its open-source bug bounty program because AI-assisted security research was generating vulnerabilities at a much faster pace. The increase in reports has also led to “triage fatigue,” where maintainers spend significant time reviewing large numbers of reports, including some generated or hallucinated by AI systems.
Daybreak is built around three AI model variants. The standard GPT-5.5 model includes regular safeguards for general use cases. GPT-5.5 with Trusted Access for Cyber is designed for verified defensive cybersecurity work. The company has also developed GPT-5.5-Cyber, a more permissive version intended for red teaming, penetration testing, and controlled validation tasks.
Several major cybersecurity companies are already working with OpenAI under the Trusted Access for Cyber initiative. These include Cisco, Cloudflare, CrowdStrike, Fortinet, Oracle, Palo Alto Networks, and Zscaler. OpenAI says it is also working with industry and government partners to develop more advanced cyber-focused AI systems in the future.
