U.S. Lifts Export Bans on Anthropic AI Models

The U.S. Department of Commerce lifted export controls on Anthropic PBC's Claude Fable 5 and Mythos 5 AI models on July 1, 2026, ending an 18-day suspension. The Trump administration originally banned the models on June 12 after reports and tests, including Project Glasswing, revealed that the systems could be exploited to identify software vulnerabilities in critical infrastructure and classified government systems. In response, Anthropic disabled the models globally to comply with directives barring access for foreign nationals.

Following negotiations, the government partially reversed the ban on June 26, allowing Mythos 5 to be redeployed to approximately 100 trusted U.S. organizations. By July 1, the government fully withdrew the export licenses for both models. To secure this restoration, Anthropic agreed to proactively detect security risks, share information on jailbreaks, and provide the U.S. government with pre-release access for evaluation. The company also implemented new safety classifiers designed to block 99% of cybersecurity vulnerability exploits.

These actions stem from a June 2 executive order by President Donald Trump establishing a voluntary framework for the federal government to vet advanced AI models for up to 30 days before public release. This oversight extended to OpenAI, which limited the rollout of its GPT-5.6 models to a small group of approved partners. While the administration aims to protect national security, industry leaders like Sam Altman criticized the government's role in selecting customers, warning that such restrictions could hinder American AI leadership against China.