Commerce Department forces Anthropic to pull Fable 5 and Mythos 5 offline, three days after launch

An export control directive issued June 12 at 5:21pm ET ordered Anthropic to suspend all foreign-national access to its two most capable models. The company complied worldwide but publicly disputed the jailbreak evidence, warning the standard would 'essentially halt all new model deployments for all frontier model providers.'

At 5:21pm ET on June 12, the US Commerce Department sent Anthropic an export control directive ordering it to suspend all foreign-national access to Fable 5 and Mythos 5, “whether inside or outside the United States.” Three days after Fable 5’s launch, the company’s two most capable models went dark globally. Because Anthropic can’t distinguish foreign nationals from domestic users in real time, compliance meant a hard shutoff for everyone. All other Claude models remain accessible.

The trigger, per Axios reporting cited by The Next Web, was an unnamed company that told the government it had found a bypass of Fable 5. Bloomberg confirmed the Commerce letter through an unnamed US official. The directive cited national security authorities without specifics, and Anthropic says the only evidence it has received is verbal.

Anthropic’s public statement is unusually combative for a company widely expected to pursue an IPO this year. The jailbreak technique, it says, “essentially consists of asking the model to read a specific codebase and fix any software flaws.” The vulnerabilities surfaced were minor and already publicly known. GPT-5.5, the company notes, can find the same flaws without any jailbreak at all. No tester has produced a universal jailbreak. The company “disagrees that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people,” and warns that the standard, applied industry-wide, “would essentially halt all new model deployments for all frontier model providers.”

The irony, flagged by TechCrunch, is structural. Mythos 5 exists precisely because Anthropic productized its vulnerability-finding capability behind cyber safeguards, restricting that variant to vetted defenders and critical infrastructure operators. That safety-first framing, which Sam Altman in April dismissed as “fear-based marketing,” is what made the model legible to Commerce as a national security artifact in the first place. Thousands of hours of red-teaming with the US government and the UK AI Safety Institute, a 30-day retention policy adopted specifically to detect jailbreak patterns, none of it produced regulatory goodwill. It produced a regulatory category.

What Anthropic built to demonstrate seriousness is the same surface the government used to pull the plug.

Commerce Department forces Anthropic to pull Fable 5 and Mythos 5 offline, three days after launch

Sources