GPT-5.6-Cyber: Expanding Daybreak as the Cyber Defense Window Narrows

OpenAI says the model is designed for advanced authorized vulnerability research, exploit validation, and security testing through Daybreak Red. - via OpenAI