Anthropic published a statement on June 12, 2026 confirming it had been ordered to suspend access to two of its frontier models. Read the original here: anthropic.com/news/fable-mythos-access.
At 5:21pm ET, Anthropic received a US government export-control directive. Citing national security authorities, the order suspends all access to Fable 5 and Mythos 5 for any foreign national, whether inside or outside the United States, including Anthropic's own foreign-national employees. The practical result: both models had to be disabled for every customer to stay compliant. Every other Anthropic model stays online.
The cited reason is a single potential jailbreak. By Anthropic's account, the government has so far provided only verbal evidence of a narrow, non-universal exploit that amounts to asking the model to read a specific codebase and fix its software flaws. Anthropic says it reviewed the report it believes the directive is based on and validated that the same level of capability is widely available from other models, including OpenAI's GPT-5.5, and is used every day by the defenders who keep systems safe.
Anthropic is complying, but it disagrees with the decision. Its argument: recalling a commercial model deployed to hundreds of millions of people over a narrow potential jailbreak is the wrong bar. If that standard were applied across the industry, the company says it would essentially halt all new model deployments for every frontier provider.
The company is also clear about what it has not seen. It says it never received a disclosure of a concerning non-universal jailbreak that led to a harmful result. The potential jailbreaks shared with it were either entirely benign responses or minor findings that gave no Mythos-specific uplift. Anthropic ran thousands of hours of red-teaming on these models with the government and third parties, and no universal jailbreak was found.
The two models are the headline, but the precedent is the real story. A single directive took frontier models offline for hundreds of millions of users overnight, over a capability the company says is already everywhere. Anthropic says it will share more details over the next 24 hours.