Anthropic restricts access to Claude Mythos for EU countries

Anthropic has blocked access to its most powerful cybersecurity model, Claude Mythos, for EU regulators and companies. While US corporations like JPMorganChase and Apple are already using Mythos to identify vulnerabilities, European banks and government agencies are left out.

The situation is complicated by OpenAI’s decision to provide access to its ChatGPT 5.5 Cyber model to European authorities. The restriction on Mythos has raised serious concerns in the European Parliament, with lawmakers stating that it threatens Europe’s digital independence by forcing the region to rely on decisions made in Washington.

We previously reported that Mythos is already helping US banks patch critical security holes and researchers use it to find vulnerabilities in macOS. Now, European organizations cannot test their systems against this tool, creating a risk of inequality in cybersecurity.