OpenAI Expands Daybreak to Help Defenders Patch Flaws

OPENAI has expanded its cyber-defense program, Daybreak, which aims to streamline patch automation for software vulnerabilities. The newly released GPT-5.5-Cyber model has shown improved capabilities in identifying and fixing flaws, scoring a record 85.6% in vulnerability reproduction tests. Over 30 million code commits have been scanned, resulting in 500,000 identified issues. The initiative includes partnerships for open-source projects and offers limited access to the model for verified defenders. OpenAI's efforts focus on maintaining human oversight while enhancing security measures for organizations.