OpenAI releases GPT-5.4-Cyber for vetted security teams, scaling Trusted Access programme

OpenAI announced the release of GPT-5.4-Cyber, a model fine-tuned for defensive cybersecurity applications. The company stated the model features lowered refusal boundaries and binary reverse engineering capabilities, and it is scaling its Trusted Access for Cyber programme to thousands of verified defenders.

The announcement explicitly positions this move as a direct response to Anthropic's Project Glasswing announcement last week, which limited access to approximately 40 corporations. OpenAI described this as representing a philosophical split between broad verified access and tightly gated deployment approaches to advanced AI models.

The model is specifically designed for cybersecurity professionals and includes capabilities tailored to defensive security work. OpenAI's Trusted Access programme expansion aims to make the model available to a larger number of verified users in the security industry.