OpenAI has revealed plans to release its latest AI model, GPT-5.4-Cyber, in a limited capacity to specific companies and users. This specialized model is designed for cybersecurity applications and is said to have enhanced capabilities in identifying security vulnerabilities compared to its predecessors.
The unveiling of GPT-5.4-Cyber follows Anthropic’s recent introduction of Claude Mythos to a select group of companies like Amazon and Microsoft. Anthropic claims that Claude Mythos has uncovered numerous zero-day cybersecurity weaknesses.
According to a recent blog post by OpenAI, GPT-5.4-Cyber is a customized iteration of the existing GPT-5.4 model. The company states that this version facilitates the identification of cybersecurity flaws, a task that the standard GPT-5.4 might decline to perform. This specialized model enables cybersecurity professionals to assess software for potential malware and vulnerabilities without requiring access to the source code.
While GPT-5.4-Cyber may appear similar to Claude Mythos at first glance, there are key distinctions between the two. GPT-5.4-Cyber is a modification of the GPT-5.4 model, whereas Claude Mythos represents a new generation for Claude, distinct from any existing version.
OpenAI’s Trusted Access for Cyber program, launched in February 2026, encompasses the cyber model, enabling specific customers and cybersecurity experts to evaluate the organization’s cutting-edge technologies.
OpenAI has specified that GPT-5.4-Cyber will initially be available to a limited group of users through the cyber program. Interested individuals or companies can request access after undergoing a verification process to join the Trusted Access for Cyber program. To gain entry, individuals must verify their government ID and complete additional authentication steps. Companies can also request access through their designated OpenAI representative.
The introduction of GPT-5.4-Cyber reflects a broader trend in the AI sector towards models dedicated to cybersecurity. The emergence of Claude Mythos has raised concerns in financial and governmental sectors after identifying numerous security vulnerabilities undetected by humans. This development has spurred discussions led by US Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell, with the Treasury Department seeking access to Mythos to bolster vulnerability identification efforts.