Anthropic’s co-founder said the company is in talks with the Trump administration about its advanced AI model, Mythos, despite tensions with the Pentagon. The U.S. Department of Defense recently cut ties with Anthropic after a disagreement over restrictions on how the military could use its AI systems, even labeling the company a supply-chain risk and banning its use by the Pentagon and its contractors.

Speaking at an event in Washington, co-founder Jack Clark emphasized that, despite the dispute, Anthropic remains committed to national security and believes the government should stay informed about its technology. He confirmed ongoing discussions about Mythos and future AI models, though details about which government agencies are involved have not been disclosed.

Mythos, introduced on April 7, is described as the company’s most advanced model for coding and autonomous tasks. Experts note that its strong coding abilities could allow it to uncover cybersecurity weaknesses and potentially exploit them. A federal appeals court in Washington, D.C., recently allowed the Pentagon’s blacklist of Anthropic to remain in place for now, marking a legal win for the Trump administration after conflicting rulings in related cases.