US lifts export controls on Anthropic's Claude Mythos, once flagged as national security threat

The US government has lifted export controls on Claude Mythos, Anthropic's top-tier AI model that Washington had blocked from export on national security grounds.

Anthropic said Monday that the Commerce Department notified it of the removal of export control directives on Mythos 5 and Fable 5, 18 days after they were imposed. "We plan to resume service starting Tuesday, July 1," the company said. "We are grateful to our users for their patience and to everyone who worked to redeploy the models."

Mythos 5 drew widespread alarm in the security industry for its ability to detect cybersecurity vulnerabilities at an expert level. Fable 5 is a version of Mythos 5 with additional safety guardrails — it is designed to refuse responses on sensitive topics that could be exploited for hacking or weapons manufacturing.

The Commerce Department imposed the export controls on June 12, citing national security risks posed by Mythos 5 and Fable 5. Because Anthropic had no way to verify users' nationalities, it blocked access to both models for all users, including Americans. On June 26, the US government partially eased the restrictions, allowing some 100 domestic companies and institutions to use Mythos.

The government imposed the controls after determining that users could bypass Anthropic's safety measures through so-called "jailbreaking" techniques.

As AI models rapidly expand their reach, governments are broadening their efforts to regulate them in the name of industrial and national security. When OpenAI unveiled its new AI model GPT-5.6 on June 26, it gave early access to pre-vetted "trusted partners" coordinated with the government — a process that amounts to a form of prior review. OpenAI has said it does not believe "this type of government approval process should become a long-term standard."