OpenAI Expands Cyber Defense Program with Tiered Access to GPT-5.4-Cyber
OpenAI is adding tiers to Trusted Access for Cyber, with top-tier defenders able to request access to GPT-5.4-Cyber.
OpenAI is expanding its Trusted Access for Cyber program with additional tiers for authenticated cybersecurity defenders. Customers at the highest tiers can now request access to GPT-5.4-Cyber, a version of GPT-5.4 fine-tuned for cybersecurity use cases and built to support more advanced defensive workflows. The company announced the expansion on X.
OpenAI has built the program around three stated principles: democratized access, iterative deployment, and ecosystem resilience. The tiered structure reflects a deliberate scaling approach — as model capabilities advance, access expands for verified defenders rather than contracting. For security teams and platform builders working on threat detection, incident response, or vulnerability analysis, a purpose-built fine-tuned model in this tier structure represents a concrete shift from general-purpose AI use toward specialized, credentialed access.