Technology

Anthropic suspends Fable 5 and Mythos 5 after US export order

Washington shut down Anthropic’s Fable 5 and Mythos 5 over a suspected jailbreak, then reversed course 18 days later after a partial thaw for 100-plus trusted users.

Sarah Chen··2 min read
Published
Listen to this article0:00 min
Anthropic suspends Fable 5 and Mythos 5 after US export order
Photo illustration

The Commerce Department lifted export controls on Anthropic’s Fable 5 and Mythos 5 on June 30, ending a suspension that had forced the company to disable the models worldwide after a June 12 directive reached Anthropic at 5:21 p.m. ET. The order required the company to stop access for any foreign national, including foreign national Anthropic employees, whether inside or outside the United States.

Officials did not spell out the national-security concern in their letter, but the company understood that U.S. authorities believed they had found a way to bypass, or jailbreak, Fable 5. The shutdown took effect the same day as the directive, leaving the two models offline just days after their June 9 launch.

AI-generated illustration
AI-generated illustration

Both systems were tightly controlled from the start. The models were red-teamed for thousands of hours before release, with testing involving the U.S. government, the UK AI Security Institute, private third-party organizations and internal teams. Mythos 5 was initially deployed through Project Glasswing in collaboration with the U.S. government for cyberdefense uses, and Anthropic set pricing for the models at $10 per million input tokens and $50 per million output tokens.

Anthropic backs tougher controls on advanced chips and model weights, including the Commerce Department’s 2025 diffusion framework, tying America’s compute advantage to both national security and economic competitiveness. The company has also warned that illicit model distillation can strip away safeguards and allow foreign labs to feed unprotected AI into military, intelligence, surveillance and offensive cyber systems.

A high-confidence Chinese state-sponsored group used Claude Code in the first documented large-scale cyberattack executed without substantial human intervention. The attackers tried to infiltrate roughly thirty global targets, including major technology companies, financial institutions, chemical manufacturers and government agencies.

The June 30 reversal came after an earlier partial easing on June 26, when officials allowed Mythos 5 to return to more than 100 trusted organizations, including groups involved in critical infrastructure defense. Access would begin being restored on July 1.

This article was produced by Prism’s automated news system from verified source data, official records, and press releases, then run through automated quality and moderation checks before publishing. The system is built and supervised by the people who set the standards it runs under. Read our full AI policy.

Did this article answer your question?

Discussion

More in Technology