US government shuts down Anthropic's Claude Fable 5 and Mythos 5 over security fears

On Friday, the U.S. government ordered Anthropic to immediately shut off access to two of its most powerful AI models — Claude Fable 5 and Claude Mythos 5 — citing national security concerns. Anthropic announced on X that it has complied, but made clear it believes the government overreacted. The directive forces the company to disable both models for all users worldwide, not just the foreign nationals the export control order nominally targets.

Mythos, Anthropic's most capable model, was kept under tight restriction since early April due to its exceptional ability to find security vulnerabilities across all major operating systems and browsers. It was shared only with about 50 vetted organizations like Amazon, Apple, and Google through Project Glasswing. Fable 5, released just days ago, was a guardrailed version meant for public use. Anthropic argues the alleged jailbreak is narrow and that similar capabilities exist in models like GPT-5.5. The company warns that if this standard applies industry-wide, it would halt all new model deployments. The shutdown comes as Anthropic pursues an IPO, and critics like Sam Altman have called Anthropic's safety messaging "fear-based marketing."