OpenAI has launched Daybreak, a brand new cybersecurity platform powered by GPT-5.5 and the Codex agent framework. Dawn is designed to assist tech corporations establish and repair software program vulnerabilities, and is a direct response to Anthropic’s Undertaking Glasswing.
The platform goals to embed cyber protection into software program from the start of the event course of, not simply as an afterthought. Dawn builds on the success of OpenAI’s earlier GPT-5.4-Cyber mannequin, which helped remediate over 3,000 safety flaws.
It makes use of Codex Safety to create an editable risk mannequin from an organization’s software program repository and focuses on essentially the most real looking assault paths. It additionally validates fixes in an remoted atmosphere, guaranteeing patches work with out inflicting unintended points.
The initiative positions OpenAI as a direct competitor to Anthropic’s Undertaking Glasswing, which main tech companies like Apple, Google, Microsoft, and Amazon have already adopted. OpenAI CEO Sam Altman mentioned the corporate desires to work with as many organizations as potential to assist them constantly safe their software program. Three GPT-5.5 variants can be found for various use instances, with stronger verification and entry controls for specialised workflows.
As cybersecurity threats develop extra subtle, Dawn’s proactive strategy is searching for to redefine how software program is developed and maintained. OpenAI is already working with business and authorities companions to arrange for the following wave of cyber-capable fashions, signaling a long-term dedication to AI-driven safety options.
With AI’s steady improvement and seemingly infinite sophistication, safety capabilities should maintain tempo with AI’s intelligence. Anthropic and OpenAI are each trying to calm that worry with their respective releases.