What happened
OpenAI released GPT-5.6-Cyber, a model built for cybersecurity tasks. The model is available through the Daybreak Red program. It supports authorized vulnerability research, exploit validation, and security testing.
The context
Cybersecurity testing relies on tools that can analyze code for flaws. OpenAI limits access to this specialized model through its Daybreak Red platform.
Sources
- Official source, https://openai.com/index/expanding-daybreak-as-the-cyber-defense-window-narrows Primary source