Summary
OpenAI has launched GPT-5.6-Cyber, a specialized AI model designed for advanced cybersecurity tasks like vulnerability research and exploit development, with significantly reduced refusals compared to general-purpose models. This fine-tuned version of GPT-5.6 Sol is exclusively available to approved organizations through OpenAI's new Daybreak Red program, which requires stringent security controls and a demonstrated need for defensive cyber work.