OpenAI expands Daybreak cybersecurity initiative as AI agent threats evolve

OpenAI has introduced two new tiers of its Daybreak service: Daybreak Blue, which offers access to frontier models, and Daybreak Red, which provides purpose-trained cybersecurity models. According to OpenAI, Daybreak Red includes GPT-5.6-Cyber, a cybersecurity-specific model.

The company has expressed concerns about the potential capabilities of an upcoming AI model called Astra. OpenAI stated that it cannot rule out that Astra has reached a "critical" capability level, meaning it could potentially launch cyberattacks against sophisticated cyber defenses autonomously.

This concern led OpenAI to pause some internal development and trigger safety protocols. The company's safety guidelines define a "critical" threshold as the ability to autonomously identify and exploit severe, real-world software vulnerabilities or execute complex cyberattacks against highly secure targets without human intervention, as per The Hindu - Sci-Tech.

Brief written by urgent.news from Techmeme, CNBC, CNBC World, CNBC Technology, The Hindu - Sci-Tech — 5 reports on this story. Machine-written — may contain errors; check the original before relying on it.