Anthropic's most advanced AI models are back online after an unusual intervention by the U.S. government. According to Anthropic, U.S. authorities have lifted a ban and export controls that had halted access to its Claude Fable 5 and Mythos 5 models, allowing the company to re-release them.
The outage was not brief. MarketScale describes it as a 19-day shutdown, framing the episode as a lesson for enterprises that have come to treat AI as core infrastructure. Digital Trends reports that Fable 5 is returning with a temporary usage limit, so access is not immediately back to normal.
The terms of the reversal offer a window into how governments are now shaping AI releases. According to the Financial Times, a U.S. government letter states that Anthropic "agreed to proactively detect and address security risks" of Fable 5 and Mythos 5, and a source said the company developed a "safeguard." The American Bazaar attributes the reversal of restrictions to the Trump administration.
The restoration is not universal. CX Today reports that access is back "but not for everyone," and Bloomberg reports that continuing U.S. limits on Mythos are keeping foreign firms in limbo. American Banker notes that questions still linger about the arrangement.
Commentators see a larger shift. TechRadar argues that while Fable 5 is back, its shutdown signals a future in which advanced AI may effectively be government-approved before it can reach users. Nextgov/FCW and MediaNama similarly frame the move as the government easing export controls on key models.
Why it matters: the episode shows that access to frontier AI can now hinge on direct negotiations between a company and the state, meaning the models businesses and consumers rely on may increasingly be gated by government approval.