GPT-5.5-Cyber
gpt-5.5-cyber
ProprietaryMultimodalCybersecurity
About
GPT-5.5-Cyber is a restricted-access cybersecurity-oriented deployment of GPT-5.5 reported on April 30, 2026 and tracked as a TAC variant because LLMReference already models GPT-5.4-Cyber separately. OpenAI's GPT-5.5 release notes state that Trusted Access for Cyber gives verified users expanded access to the advanced cybersecurity capabilities of GPT-5.5 with fewer restrictions, starting with Codex, and that users can apply at chatgpt.com/cyber for verified defensive work. There is no public API model ID or self-serve endpoint, and OpenAI has not published separate context, pricing, or parameter details for this restricted cyber access path.
Capabilities
VisionMultimodalReasoningFunction CallingTool UseStructured OutputsCode ExecutionPrompt CachingBatch APIAudioFine-tuning
