Days after rival Anthropic launched Mythos, OpenAI announces GPT-5.4-Cyber AI model built for cybersecurity companies

ChatGPT-maker OpenAI has launched GPT-5.4-Cyber, a variant of its flagship model GPT 5.4 launched in March this year. The new GPT-5.4-Cyber is specifically designed for defensive cybersecurity work. Announcing the new model, OpenAI said: "This is a version of GPT‑5.4 which lowers the refusal boundary for legitimate cybersecurity work and enables new capabilities for advanced defensive workflows, including binary reverse engineering capabilities that enable security professionals to analyze compiled software for malware potential, vulnerabilities and security robustness without needing access to its source code". Notably, the launch comes days after OpenAI-rival Anthropic introduced its frontier AI model Mythos.Similar to Anthropic which said that it will release the mythos model to select companies, OpenAI's GPT-5.4-Cyber will initially be available to vetted security vendors, organizations and researchers because of its more permissive design. "Because this model is more permissive, we are starting with a limited, iterative deployment to vetted security vendors, organizations, and researchers. Access to permissive and cyber-capable models may come with limitations, especially around no-visibility uses like Zero-Data Retention⁠(opens in a new window) (ZDR)," the company said.Additionally, the company is expanding its Trusted Access for Cyber⁠ (TAC) to bring additional tiers of access for users willing to work with OpenAI to authenticate themselves as cybersecurity defenders. For those unaware, TAC was introduced in February this year with both automated identity verification for individuals in order to reduce the friction of safeguards on cybersecurity-related tasks and partner with a limited set of organizations for more cyber-permissive models.OpenAI said that customers in the highest tiers will get access to GPT‑5.4‑Cyber. How to gain access to OpenAI's TAC As announced by the company, following users can get access to TAC:GPT-5.4-Cyber is based on an existing GPT model, while Mythos is a completely new version of Claude. Anthropic plans to offer Mythos to a limited number of companies. It is part of the company's "Project Glasswing", a controlled initiative under which select organizations are permitted to use the unreleased ⁠Claude Mythos Preview model for defensive cybersecurity purposesOpenAI's GPT-5.4-Cyber model is part of a controlled access programme for select users and cybersecurity experts.