OpenAI Launches Daybreak Program to Automate Cyber Defense

OpenAI launched the Daybreak cybersecurity initiative to transition AI-driven security from simple vulnerability discovery to active remediation. Central to this effort is the full release of GPT-5.5-Cyber, a specialized model for vetted defenders that outperforms standard versions in reproducing vulnerabilities. The initiative includes an updated Codex Security plugin for automated patch generation, which has already scanned 30,000 repositories and assisted in fixing over 500,000 findings.

As part of the effort, OpenAI established the Daybreak Cyber Partner Program, granting frontier AI access to a curated group of defenders. IBM joined the program to launch an application security service, supported by a $5 billion commitment from IBM and Red Hat to Project Lightwell for securing open-source code. Other partners include TrendAI, which integrated agentic triage into its Vision One platform, as well as Tenable Holdings, Inc., Proofpoint, Inc., CrowdStrike, Sophos, and Fortinet.

To support the open-source community, OpenAI co-founded the Patch the Planet initiative with Trail of Bits and HackerOne. This program pairs security researchers with maintainers of projects like Python, Go, and cURL to convert AI-identified vulnerabilities into verified fixes. In its first week, the collaboration identified hundreds of bugs across 19 projects. OpenAI is also extending these defensive standards to government agencies and critical infrastructure operators in the United States, Japan, and the European Union.