Anthropic Suspends Claude Fable 5 and Mythos 5 After US Government Order
Anthropic disabled access to its Claude Fable 5 and Mythos 5 models for every customer after the U.S. government ordered the company to bar foreign nationals from using the tools, citing undisclosed national security concerns. Because the restriction could not be enforced cleanly across a global user base, the company took the broader step of suspending both models worldwide. For anyone who had built a workflow around Fable 5, Anthropic’s most capable public model and a free-in-beta sensation, the sudden outage is a sharp lesson in why no AI stack should depend on a single model.
What happened?
On June 13, 2026, Anthropic posted a statement confirming it had received an order from the U.S. Department of Commerce prohibiting foreign nationals from using Claude Fable 5 and Mythos 5. Unable to enforce that restriction technically across its worldwide user base, the company disabled access for all customers, effective immediately.
The government did not publicly detail the threat. Anthropic said it believed the trigger was a demonstration of a jailbreaking technique that could make the model identify previously known software vulnerabilities. After reviewing the technique, the company said the vulnerabilities it surfaced were minor and that other publicly available AI models could find them without any bypass. The order forced a complete takedown regardless.
“The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.” Anthropic, in its official statement suspending the models.
How do the two models fit together?
Fable 5 is the production version of Anthropic’s Mythos architecture, the same family the company had earlier described as too powerful to release. Ahead of the public launch, Fable 5 had been made available free through June 22, which drove rapid adoption among creators, agencies, and teams who praised its long-context reasoning and creative fluency. Mythos 5 served as the enterprise-facing counterpart. Both were taken offline at once.
The suspension does not affect older Claude models, which remain available and unaffected. But it underscores a pattern. Anthropic is already locked in a separate dispute with the Trump administration after the Pentagon labeled the company a supply chain risk and tried to bar agencies from using its AI. A federal judge temporarily blocked that directive, but the broader regulatory friction is far from settled.
What are the key facts?
- Two flagship models disabled. Claude Fable 5, the public-facing model, and Mythos 5, the enterprise counterpart, were both taken offline, affecting every Anthropic customer.
- Free access had accelerated adoption. Fable 5 was free through June 22, prompting widespread testing across creative and technical teams.
- A jailbreak demonstration triggered the order. The government was shown a bypass that extracted minor, previously known vulnerabilities. Anthropic noted other models could find the same flaws without any bypass.
- Pre-release caution was already in place. Anthropic had allowed a small group of organizations to preview Mythos in April specifically to probe its cybersecurity risks, acknowledging the model’s exploit abilities warranted caution before a wide release.
What comes next?
Anthropic said it is working with authorities to understand the order’s exact scope and explore ways to restore access, with no timeline shared. The European Union, which had only just gained access to Mythos after weeks of negotiations, called the development further proof that the bloc needs technological sovereignty, a signal that non-U.S. governments may accelerate their own AI infrastructure efforts.
Meanwhile, the administration’s broader posture toward AI companies it views as security risks is intensifying. The outcome of Anthropic’s lawsuit over the supply chain risk label could set a precedent for how far the government can go in throttling civilian access to advanced AI tools.
What is the bigger picture?
This is more than one company’s compliance scramble. It is a signal that frontier AI models are now treated as dual-use technologies, and that governments will intervene quickly, sometimes without a clear public explanation, when they decide a tool is too capable. For anyone who depends on AI in production, the practical lesson is direct: a model you built a workflow around can vanish overnight. The teams that weather these shocks are the ones who diversify their providers, keep a tested fallback model ready, and treat no single engine as irreplaceable. Build the stack so that a government order, a policy reversal, or a jailbreak headline does not take your work down with it.
FAQ
Why did Anthropic suspend Claude Fable 5 and Mythos 5?
Anthropic suspended both models after receiving an order from the U.S. Department of Commerce prohibiting foreign nationals from using them. Because the company could not enforce that restriction across its global user base, it disabled access for all customers.
What triggered the government order?
Anthropic said the trigger was a demonstration of a jailbreaking technique that could make the model identify previously known software vulnerabilities. The company reviewed the technique and found the surfaced vulnerabilities were minor and discoverable by other public AI models without any bypass.
Will Fable 5 or Mythos 5 come back online?
Anthropic said it is working with authorities to understand the order’s exact scope and explore ways to restore access, but no timeline has been shared. Older Claude models remain available and are unaffected by the suspension.