OpenAI begins rolling out GPT-6 Cyber to select Daybreak Red members

OpenAI is launching GPT-6 Cyber to select members of its Daybreak Red program, Emma Roth reports for The Verge. The program is available to organizations using OpenAI’s models for authorized defensive cybersecurity work. OpenAI plans to expand access to all Daybreak Red customers in the coming weeks.

The model carries a Critical cybersecurity classification under OpenAI’s Preparedness Framework. As The Verge reports, that designation indicates a model can develop functional zero-day exploits in many hardened critical systems without human intervention. Its predecessor, GPT-5.6 Cyber, was classified High. The classification describes OpenAI’s assessment, not an independently demonstrated account of GPT-6 Cyber’s capabilities.

The Verge reports that on ExploitGym, a benchmark measuring whether models can exploit vulnerabilities, GPT-6 Cyber scored 48.5 percent. GPT-5.6 Cyber scored 32.1 percent, and GPT-6 Astra scored 42.4 percent. OpenAI also credits the new model with helping find 81 previously unknown Firefox vulnerabilities and hardening its AI training infrastructure.