EU Cybersecurity Agency Tests Anthropic's Mythos 5 AI

The European Union Agency for Cybersecurity (ENISA) has gained access to Anthropic’s powerful AI model, Mythos 5, and is currently testing it, according to the European Commission. This development marks a significant step in the EU's efforts to evaluate advanced artificial intelligence systems for their cybersecurity implications. Mythos 5 has been subject to restricted access since its release in April 2026 due to its advanced capabilities in identifying and exploiting software vulnerabilities. Anthropic initially made the model available to a select group of partners through its Project Glasswing, which began with approximately 50 organizations and later expanded to over 200. The European Commission confirmed ENISA's access on Thursday, September 10, 2026, following months of negotiations. Negotiations for ENISA's access were reportedly complicated by US export controls on advanced AI technologies, which temporarily restricted foreign access to models like Mythos 5 and Fable 5. These controls, which remained in place for about two weeks, led to friction and prompted questions from the European Commission regarding potential discrimination against trusted partners.