NSA loses access to Anthropic’s Mythos 5 and Fable 5 AI models after Trump administration export controls

The U.S. National Security Agency has been forced to stop using Anthropic’s latest AI models, Mythos 5 and Fable 5, after the Trump administration imposed export controls on the startup. The models had shown unusually strong capability in internal NSA testing to uncover software vulnerabilities. In a congressional hearing this month, Sen. Mark Warner said NSA Director Gen. Joshua Rudd told him Mythos broke into “almost all of our classified systems” within hours, comments later cited by The Economist.