OpenAI has introduced a new AI model, with some outlets specifying it as GPT-5.4-Cyber, featuring advanced capabilities for security teams, including reverse engineering, vulnerability detection, and malware analysis. The model appears to be part of an expanded cyber defense program, although the exact version number is unclear due to a conflicting report of a GPT-5.5 introduction. The new model is reportedly tailored for defensive cybersecurity and is available to vetted researchers.