U.S. Export Controls Force Anthropic to Halt Mythos 5 and Fable 5 Models
The U. S. Commerce Department placed Anthropic’s Claude Mythos 5 and Claude Fable 5 AI models under export controls, forcing the company to block foreign-national use. WIRED reported that Anthropic reversed a covert performance-degradation policy for Claude Fable 5 after researchers criticized the hidden safeguard and said the safeguards would be made visible. TechCrunch and Ars Technica later reported that the Commerce Department issued the export-control directive citing national-security risks and a reported jailbreak, requiring Anthropic to disable both models worldwide. Anthropic complied, halting access for foreign nationals while keeping smaller Claude models available, according to Fortune and the Los Angeles Times.
Business Insider added that the Trump administration imposed the restrictions after high-level meetings over security concerns. Fortune noted that the Commerce Department halted distribution after Amazon identified a security breach, marking the first export-control action against a commercial AI system. The Washington Post said the White House gave Anthropic a 90-minute window to shut down the newly launched Fable model. TechCrunch later reported that cybersecurity veterans urged the government to lift the limits, arguing they hinder defenders. WIRED and The Verge documented ongoing talks in Washington, with Anthropic and the administration remaining at odds over the export controls, leading to continued negotiations.