OpenAI launches GPT-5.4-Cyber days after Anthropic's Mythos reveal

OpenAI has unveiled GPT-5.4-Cyber, a specialised version of its latest flagship model tailored for defensive cybersecurity applications, just days after rival Anthropic announced its own frontier AI model, Mythos, according to a report by Reuters.

Anthropic had introduced Mythos on April 7 as part of its Project Glasswing initiative, a controlled programme that allows select organisations to access the unreleased Claude Mythos Preview model for defensive cybersecurity use. The model has reportedly identified thousands of major vulnerabilities across operating systems, web browsers and other software systems.

OpenAI, the company behind ChatGPT, stated that GPT-5.4-Cyber will initially be deployed on a limited basis to vetted security vendors, organisations and researchers, citing its more permissive design compared to standard models.

The company also informed that it is expanding its Trusted Access for Cyber (TAC) programme to include thousands of verified individual defenders and hundreds of teams working to secure critical software infrastructure, as outlined in a post on its website.

As part of this expansion, OpenAI is introducing additional tiers within the TAC programme, which was first launched in February. Higher levels of verification will unlock access to more advanced capabilities.

Users granted approval under the highest tier will gain access to GPT-5.4-Cyber, which operates with fewer restrictions on sensitive cybersecurity tasks, including vulnerability research and analysis.