The National Security Agency (NSA) is using Anthropic's most powerful model, Mythos Preview, despite the Department of Defense (DOD) claiming it is a supply chain risk. The Pentagon moved in February to cut off Anthropic and force its vendors to follow suit. The military is expanding its use of these tools while arguing in court that they threaten U.S. national security. The relationship between the Pentagon and Anthropic began during tense contract renegotiations earlier this year, with the department demanding that Anthropic make its Claude model available for "all lawful purposes", while the company insists on walling off mass domestic surveillance and the development of autonomous weapons. The NSA's counterparts in the UK have also said they have access to the model through the country's AI Security Institute.