US Lifts Anthropic Export Control Restrictions, AI Models Return Online

The US has lifted export control restrictions on Anthropic's AI models, allowing them to return online. Anthropic has confirmed that Claude Fable 5 and Mythos 5 have been made available globally on the Claude Platform, and will be included for up to 50% of weekly usage limits through July 7. The decision is seen as a win for Amazon and Anthropic.

The US Department of Commerce has lifted export controls on Anthropic's most advanced AI models, clearing Claude Fable 5 and Mythos 5 to return online after roughly three weeks offline . Anthropic said it would begin restoring global access to the models the following day, reversing an order that had required the company to block foreign nationals, including some of its own employees, from using them.

The original restriction followed claims by Amazon researchers that certain prompts could coax Fable 5 into producing information useful for cyberattacks. Over the following two weeks, Commerce worked with Anthropic to analyze and re-approve the models, removing the licensing requirements that had applied to their export and transfer. The episode underscores how national-security review is becoming a gating factor for frontier-model deployment.

The reversal is a clear positive for Amazon AMZN, Anthropic's largest strategic investor and cloud partner, which benefits from renewed model availability on its infrastructure. It also lands as Anthropic weighs a potential IPO and faces intensifying competition, including fast-improving Chinese models that have narrowed the gap with US labs. Investors will watch whether clearer export rules reduce the risk of future shutdowns for US frontier developers.