Reports indicate that the National Security Agency is utilizing Mythos Preview, a highly advanced cybersecurity model developed by Anthropic that remains restricted from public release. This revelation emerges shortly after the Department of Defense, the parent agency of the NSA, classified Anthropic as a supply chain risk. The designation followed a dispute where the AI firm refused to grant Pentagon officials unrestricted access to the full capabilities of its models. Anthropic introduced Mythos earlier this month as a frontier model specifically engineered for cybersecurity operations. However, the company determined that the system possessed sufficient power to facilitate offensive cyberattacks if released openly. Consequently, access has been limited to approximately 40 organizations, with only a dozen publicly named. The NSA is now understood to be one of the undisclosed recipients, deploying the model primarily to scan digital environments for exploitable vulnerabilities. The UK AI Security Institute has also confirmed its access to the system. The situation presents a complex contradiction: the U.S. military is expanding its adoption of Anthropic tools while simultaneously litigating that such technology poses threats to national security. The initial rift between the Pentagon and Anthropic began when the company declined requests to make its models available for mass domestic surveillance and autonomous weapons development. Meanwhile, relations between Anthropic and the Trump administration appear to be improving. Last Friday, CEO Dario Amodei held a meeting with White House Chief of Staff Susie Wiles and Treasury Secretary Scott Bessent. Sources describe the discussion as productive. This shift suggests a potential softening of the previously adversarial stance between the agency and the tech company. TechCrunch sought comment from the NSA regarding these reports, but the agency has not responded. Anthropic declined to provide a statement on the matter. The convergence of the NSA using the restricted model while the Department of Defense previously flagged the firm as a risk highlights the intricate and often conflicting dynamics between national security interests and the deployment of advanced artificial intelligence. The situation underscores the delicate balance between leveraging cutting-edge technology for defense and managing the inherent risks associated with dual-use AI capabilities. As the government continues to navigate these challenges, the selective distribution of models like Mythos remains a critical factor in the broader conversation regarding AI safety and national defense strategy.