Prime Highlights :
- OpenAI warned that AI-driven cyberattacks will grow faster, larger, and more autonomous, leaving defenders a shrinking window to prepare.
- Blue was named the recommended starting point for most defenders, while Red offers advanced tools for security testing and vulnerability research.
Key Facts :
- Daybreak is OpenAI’s cyber defense service, now split into Blue and Red tiers for approved customers.
- GPT-5.6-Cyber, exclusive to the Red tier, is currently limited to select partners like Accenture, IBM, Crowdstrike, and Cloudflare.
Background :
OpenAI has expanded its cyber defense service Daybreak, introducing a new cybersecurity model built specifically for defensive work. The announcement came in the second week of August, months after the company first launched Daybreak earlier this year.
The expansion arrives as reports of AI systems behaving unpredictably have grown more frequent. Some recent examples are a data breach in a large AI platform, an AI agent compromising a gym website, and the use of AI systems to create fictitious profiles for conducting social engineering attacks.
Daybreak now offers two service tiers, Blue and Red, both giving approved customers access to OpenAI’s limited-access frontier cyber models. Blue provides a range of defensive tools, including incident response, malware analysis, and patch validation. OpenAI described Blue as its recommended starting point for most organizations seeking protection.
Red offers a wider set of tools meant for security testing and vulnerability research. This tier includes access to the newly launched GPT-5.6-Cyber model, built on the company’s GPT-5.6 Sol system. The model is currently available only to select trusted partners, reportedly including Accenture, IBM, CrowdStrike, and Cloudflare.
OpenAI stated that the cybersecurity landscape is changing quickly, noting that threat actors are expected to increasingly use AI to launch attacks at greater speed and scale, including in fully autonomous ways. The company added that defenders now have a shrinking window of time to prepare.
Industry observers have noted that such expansions also serve as marketing tools for AI companies. Still, many enterprises continue turning to AI labs for protection, given their direct experience with the risks these systems create.