OpenAI launched its Daybreak cybersecurity initiative on May 12, 2026, introducing two AI models running on GPT-5.5: Daybreak Blue for defensive security and Daybreak Red for authorized penetration testing. Daybreak Red has already helped uncover two zero-day vulnerabilities in Google’s V8 JavaScript engine that powers Chrome. On CyberGym benchmarks, the GPT-5.5-Cyber model scored 85.6% compared to 81.8% for standard GPT-5.5. On June 22, 2026, OpenAI expanded Daybreak with the « Patch the Planet » program, developed in partnership with Trail of Bits to support open-source project maintainers.
Source: Read the original article
