Anthropic has decided to provide the European Union with access to its high-performance AI model Mythos, which it had allowed only limited access to for some organisations, CNBC reported on Sunday.

The EU has sought access to Mythos for months, citing cybersecurity concerns, and talks have recently advanced.

Mythos is a model Anthropic disclosed on a limited basis to some companies in April as part of its cybersecurity initiative Project Glasswing. The Mythos model showed strength in finding software security flaws and vulnerabilities, but it has also raised concerns that malicious actors could misuse it to accelerate cybercrime.

Spokesperson Regnier said Mythos was not a one-off case, that more powerful models would come to market, and that discussions were being intensified with like-minded partners, including the United States.

The conditions under which Anthropic will allow the EU access to Mythos have not yet been disclosed. The EU recently heard from Anthropic that government approval is required to provide Mythos, and it expanded talks last week with the U.S. administration to secure access.

The EU also secured access in May to OpenAI’s GPT-5.5-Cyber model.