OpenAI expands cyber access
OpenAI added tiers to its Trusted Access for Cyber program that give high‑level users access to a specialized model called GPT‑5.4‑Cyber. The move targets advanced cybersecurity workflows by gating a fine‑tuned model behind program tiers for vetted users. (x.com)
OpenAI said on April 14 it is widening its Trusted Access for Cyber program and giving top-tier vetted users access to a new model, GPT-5.4-Cyber. (openai.com) The company said the program will expand to “thousands of verified individual defenders” and “hundreds of teams” that protect critical software. OpenAI said GPT-5.4-Cyber is a fine-tuned version of GPT-5.4 built to be more permissive for defensive security work. (openai.com) Cybersecurity work often means finding weaknesses before attackers do, like stress-testing a lock before a burglar tries it. OpenAI said the new model is aimed at tasks such as vulnerability research, malware analysis, incident response, and secure code review for approved users. (openai.com) OpenAI launched Trusted Access for Cyber in February 2026 as an identity-based program for users handling high-risk, dual-use cyber tasks. At launch, the company also pledged $10 million in application programming interface credits to speed up defensive security work. (openai.com; openai.com) The expansion comes five weeks after OpenAI released GPT-5.4 on March 5, describing it as a frontier model for professional work with coding, tool use, and a context window of up to 1 million tokens. OpenAI’s GPT-5.4 system card said that model was the first general-purpose release in the GPT-5 line with mitigations for “High” cybersecurity capability. (openai.com; openai.com) OpenAI said it is preparing for “increasingly more capable models” over the next few months and wants vetted defenders to adopt those tools faster than attackers do. In a December 2025 post, the company said its internal cyber evaluations had climbed from 27% on capture-the-flag tests with GPT-5 in August 2025 to 76% with GPT-5.1-Codex-Max in November 2025. (openai.com; openai.com) The gatekeeping is the point. OpenAI said GPT-5.4-Cyber is available only through program tiers for verified defenders, and the company said it is pairing broader access with “strengthened safeguards” as model cyber capabilities rise. (openai.com) That approach puts OpenAI closer to a model now emerging across artificial intelligence labs: keep the most cyber-capable systems off the open shelf, then release them first to screened users with defensive jobs. OpenAI’s April 14 post framed GPT-5.4-Cyber as the first step in that tiered rollout. (openai.com)
