OpenAI's Daybreak expands to fight AI cyberattacks

OpenAI has expanded its Daybreak cyber defense platform with two distinct tiers—Blue and Red—each powered by advanced AI models designed to counter growing AI-driven cyber threats. The Blue tier serves as an entry point for enterprises, offering core cybersecurity services such as incident response, malware analysis, and patch validation. OpenAI positions Blue as the recommended starting point for most defenders, suggesting it meets the needs of typical enterprise security teams.

The Red tier, however, introduces a more specialized toolkit, including the newly unveiled GPT-5.6-Cyber model. Built on the foundation of GPT-5.6 Sol, this model is optimized for cybersecurity tasks like security testing and vulnerability research. Unlike Blue, Red is restricted to "trusted customer partners," including major firms like Accenture, IBM, CrowdStrike, and Cloudflare. OpenAI frames this expansion as a response to the escalating use of AI agents in cyberattacks, warning that defenders have a "narrowing window to prepare" as threats evolve.