OpenAI says it is expanding its Daybreak initiative with the release of an improved GPT-5.5-Cyber model for trusted defenders. The company announced the update on Monday, building on a Daybreak announcement made last month. OpenAI describes GPT-5.5-Cyber as its strongest model yet for finding and helping patch software vulnerabilities. It says the model can sustain deeper analysis across large codebases to help identify security issues and support remediation efforts.
Infosecurity Magazine reports that OpenAI expands Daybreak by moving to a full GPT-5.5-Cyber release aimed at helping defenders patch software flaws. Across the two reports, the emphasis is on providing cybersecurity-focused AI capabilities to authorized parties under the Daybreak framework, with the stated goal of improving the vulnerability discovery and patching workflow for defenders. Neither source provides detailed technical performance metrics, timelines for general availability, or information about which organizations qualify as “trusted defenders.”
