OpenAI has expanded Daybreak, its cybersecurity service, with a new model built specifically for defensive work. The update reflects how quickly AI-powered security tools are evolving as organizations look for stronger ways to protect digital systems.
The platform now includes two access levels: Blue and Red. Blue is positioned as the practical starting point for most security teams and includes support for incident response, malware analysis, and patch validation. Red is designed for more advanced testing, offering purpose-trained cybersecurity models for vulnerability research and deeper security assessments.
The new model, GPT-5.6-Cyber, is available only through the Red tier and is built on GPT-5.6 Sol. OpenAI says it is intended to strengthen specialized cybersecurity tasks for approved partners. Current trusted users reportedly include organizations such as Accenture, IBM, CrowdStrike, and Cloudflare.
The company says the pace of AI-driven security threats is accelerating, making defensive preparation more important than ever. At the same time, the expansion of Daybreak shows how AI labs are turning their own technical expertise into security products for enterprises.
As AI and cybersecurity continue to converge, tools like Daybreak could help shape a faster, more adaptive digital defense ecosystem in the years ahead.
