US Commerce Department blocks Anthropic AI models for foreign users

On June 12 2026 the U.S. Commerce Department directed Anthropic to immediately prevent all foreign nationals from accessing its two most advanced large‑language models, Claude Fable 5 and Mythos 5. Anthropic shut the models down within hours, making them unavailable worldwide.

The ban was lifted on June 30 after Anthropic added extensive safety guardrails, but the changes caused a sharp drop in the models’ benchmark performance, with routine tasks now often blocked. Legal scholars note this was the first use of the Export Control Reform Act of 2018 to restrict AI services, raising questions about the legality and future scope of U.S. export controls over emerging technologies. The episode has also heightened European concerns about reliance on U.S. AI providers and the need for an independent European AI capability.