U.S. Clears Anthropic Mythos 5 for Limited Release as AI Access Fight Continues
The U.S. government has partially lifted its block on Anthropic’s Claude Mythos 5 model, allowing the company to release the powerful AI system to more than 100 approved U.S. institutions, including major companies and government agencies.
The decision is a major de-escalation in a fast-moving fight between the Trump administration and one of the country’s most closely watched AI companies. Semafor reported that Commerce Secretary Howard Lutnick told Anthropic that “appropriate safeguards” were in place for certain trusted partners to access Mythos 5.
The approval is limited. Axios reported that export controls remain in place for organizations not explicitly approved by the administration and that the letter does not change restrictions on Fable 5, Anthropic’s broader consumer-facing model. The government also reserved the right to revise access if circumstances change.
Anthropic previously said the government directive forced it to suspend access to Fable 5 and Mythos 5 for foreign nationals, including foreign-national employees inside the company. Anthropic said it disagreed with the basis for the action, arguing the government had described only a narrow jailbreak concern and had not provided detailed national-security reasoning.
Subscribe free for daily political analysis they won’t broadcast. Join 110K+ readers →
The practical consequence is significant: Washington is now deciding which organizations can use one of the most advanced AI tools for cybersecurity work. That could reduce misuse risk, but it also gives federal officials direct influence over access to frontier AI capabilities.
The restriction drew visible pushback from cybersecurity experts. Cybersecurity Dive reported that a letter signed by 76 CEOs, CISOs, venture capitalists, and security researchers argued the ban removed strong tools from defenders and created market uncertainty.
Online developer discussion has also focused on whether foreign-national restrictions are technically workable and whether advanced AI is now being treated more like defense technology than ordinary software.
What happens next depends on whether the government expands approved access and whether Fable 5 is cleared for general use.
Subscribe free for daily political analysis they won’t broadcast. Join 110K+ readers →