Anthropic faces shutdown order for Claude Mythos 5 models

The directive arrived at 5:21 PM, leaving Anthropic leadership scrambling to disable systems they had spent the week promoting. By restricting access to foreign nationals regardless of location, the government’s mandate rendered the models virtually unusable for the company’s internal operations and external clients. Executives immediately pivoted to an emergency trip to Washington, DC, hoping to reverse a policy that threatens to upend the competitive landscape for American AI developers.

This crisis stems from the lineage of the models themselves. Both Mythos 5 and Fable 5 share a foundation with the Mythos Preview, a version Anthropic previously withheld from the public, citing extreme safety concerns. While Fable 5 was marketed as a hardened, safe-for-general-use alternative, reports surfaced suggesting its safety guardrails were insufficient. Those vulnerabilities gave the administration leverage to enforce a sudden crackdown, turning Anthropic’s own warnings about the danger of its technology into the justification for the federal intervention.