SAN FRANCISCO, June 22, Fable 5 and Mythos 5 have now been dark for ten days, and Anthropic is no longer being quiet about it. In a statement posted to its site, the company says the Commerce Department directive that forced the shutdown at 5:21 p.m. ET on June 12 still hasn’t been accompanied by a specified national-security rationale, and that restoration of the two models remains under discussion.

The legal mechanism is an export-control directive covering access “whether inside or outside the United States, including foreign national Anthropic employees.” Rather than partition its userbase, Anthropic disabled both models for everyone. Its other models stayed up. An unnamed U.S. official confirmed the letter to Bloomberg.

Anthropic’s reading is that the directive traces to a research paper by Amazon security researchers, first surfaced by The Wall Street Journal, describing a guardrail bypass in Fable 5. The company says the technique surfaces “a small number of previously known, minor vulnerabilities” that publicly available models including OpenAI’s GPT-5.5 can also discover. Its objection is structural: “We disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people.”

The subtext is harder to miss. In March the Pentagon designated Anthropic a “supply chain risk” after contract negotiations collapsed, a label the company is contesting in federal court. Two confrontations with the Trump administration inside a single quarter, both routed through national-security instruments rather than ordinary regulatory channels, suggest the export-control framing is the surface of a procurement dispute that didn’t resolve cleanly.

What’s notable is the posture. A frontier lab is publicly contradicting Commerce’s stated basis for a takedown while complying with it, betting that the legibility of its argument outweighs the cost of the silence.

Sources