OpenAI Ships GPT-5.6-Cyber Through Gated Daybreak Red Access Tier

OpenAI has released GPT-5.6-Cyber, a cybersecurity-specific model that gives approved defenders tools for exploit development and vulnerability research. The company positioned the launch around a shrinking window for defense. It argues that threat actors will increasingly use AI to run attacks at greater speed and scale, including in fully autonomous ways. OpenAI Gives Trusted Defenders