OpenAI Launches GPT-5.4-Cyber Model, Offering Reduced Refusal Rates for Security Professionals

Odaily Planet Daily News: OpenAI has announced the release of the GPT-5.4-Cyber model, which is fine-tuned based on GPT-5.4 for cybersecurity scenarios. Under specific conditions, the model reduces refusal rates for security-related requests to support professional operations such as binary reverse engineering. According to the announcement, the model is exclusively available through the 'Trusted Access for Cybersecurity' (TAC) program to vetted security vendors, corporate security teams, and researchers. OpenAI stated that layered mechanisms will restrict the model's usage scope, with additional constraints applied to access in low-visibility scenarios.