Anthropic Shuts Down Fable 5 and Mythos 5 Following U.S. Government Export Control Order

Blunt Government Move Halts Anthropic’s Advanced Model Rollout

According to The Verge AI, the U.S. government ordered Anthropic on Friday, June 13, to block all access to Fable 5 and Mythos 5 for foreign nationals and Anthropic employees, citing undisclosed national security risks. Anthropic complied by immediately cutting off access to both models across all customer accounts. The intervention represents an escalation in how governments regulate frontier AI—moving from advisory warnings to binding access restrictions on deployed commercial systems.

The Government’s Undisclosed Security Rationale

The order lacks transparency on its core justification. According to Anthropic’s statement to The Verge, the government “did not provide specific details of its national security concern” and instead disclosed evidence of potential vulnerabilities only through verbal briefings rather than written disclosure. This opacity creates a tension: Anthropic claims the vulnerabilities presented were “minor findings that provide no Mythos-specific uplift” and notes that comparable issues exist in competing models like GPT 5.5, suggesting the threat may not be Fable or Mythos-unique.

Anthropic emphasized in its response that it had already coordinated with the U.S. and UK governments and revised its data retention policies to detect malicious exploitation attempts. The company further stated it had “not even received a disclosure of a concerning non-universal potential jailbreak that led to a harmful result,” implying the government’s concerns may be theoretical rather than tied to observed misuse.

Access Scope and Enforcement

The blocking is comprehensive: all foreign nationals lose access regardless of location, and Anthropic’s own international workforce cannot use the models. This all-or-nothing enforcement contrasts with narrower alternatives—such as geographic restrictions, user verification requirements, or API-level monitoring. The scope suggests the government viewed partial measures as insufficient.

Why This Matters

This action signals a shift in regulatory posture toward frontier models. Rather than requiring security audits or imposing usage restrictions, the government exercised direct access control—a precedent that may influence how other regulators respond to AI safety concerns. For Anthropic, the financial impact depends on Fable and Mythos’s revenue contribution and whether customers can migrate to remaining models. For the industry, the case demonstrates that national security claims—even when undetailed—can trigger immediate compliance at the commercial layer, raising questions about how much evidentiary rigor should precede such orders.