US Export-Control Order Forces Anthropic to Pull Fable 5 and Mythos 5 Offline

Anthropic disabled both flagship models on Friday after a Commerce Department directive citing national security, severing access for every customer worldwide and raising new questions about regulatory exposure for production agents.

At 5:21 p.m. ET on Friday, a U.S. Commerce Department export-control directive landed at Anthropic. By that evening, Claude Fable 5 and Mythos 5 were dark worldwide. Because the order barred any foreign national from using the models, including Anthropic’s own non-citizen employees, a global shutdown was the only compliant option.

Fable 5 had been live for three days. It was the guardrailed public release of Mythos 5, which had been restricted under Project Glasswing to roughly 50 vetted organizations including Amazon, Apple, Google, Microsoft, and CrowdStrike. Per benchmarks from Vals AI, Fable 5 was the most capable model publicly available at launch. Claude Opus 4.8 and the rest of Anthropic’s catalog remain online.

The directive didn’t specify what national security concern triggered it. Anthropic, in its blog post, said it believes the government is responding to a claimed jailbreak, evidence for which has so far been shared verbally and described as a “potential narrow, non-universal jailbreak.” The behavior in question: prompting the model to read a codebase and identify software flaws. The same capability ships in OpenAI’s GPT-5.5 and is used daily by defensive security teams.

“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,” Anthropic wrote.

The technical case for the order looks thinner on inspection. Cybersecurity researcher Katie Moussouris, who reviewed the underlying paper (authored, per the Wall Street Journal, by Amazon researchers), wrote that the behavior “should never have triggered an export control” and “cannot meaningfully be fixed” without weakening the model for defenders. TechCrunch, citing Axios, reported that the directive was attributable in part to “personality differences” between Anthropic and the Trump administration rather than a technical finding.

That framing tracks with the broader pattern. In March, the Pentagon designated Anthropic a “supply chain risk,” barring military and defense-contractor use of Claude. Anthropic is challenging that designation in federal court; the litigation is ongoing. Friday’s takedown lands inside that fight, not outside it. For every team building production agents on a frontier US lab, the regulatory surface is no longer a hypothetical line item.

US Export-Control Order Forces Anthropic to Pull Fable 5 and Mythos 5 Offline

Sources