Anthropic disables Claude Fable 5 and Mythos 5 globally after US export order
Three days after launch, Anthropic pulled its newest models worldwide citing national security concerns.
On June 16, 2026, Anthropic disabled its newly released Claude Fable 5 and Mythos 5 AI models globally, just three days after their public debut on June 12. The move came in response to a U.S. export control order from the government, which cited national security concerns and the discovery of a potential 'jailbreak' that could bypass the models' built-in safety guardrails. Anthropic stated it suspended access for all foreign nationals worldwide to ensure full compliance with the directive.
This dramatic rollback underscores growing tensions between AI innovation and regulatory oversight. The affected models were Anthropic's most advanced to date, and their swift removal disrupts enterprises and developers who had already integrated them into workflows. The incident highlights the fragility of relying on cutting-edge AI models that can be pulled due to geopolitical or security mandates, raising urgent questions about the future of AI deployment, export controls, and the balance between capability and control.
- Claude Fable 5 and Mythos 5 were released on June 12, 2026, and disabled on June 16, 2026.
- U.S. export control order cited national security concerns and a potential jailbreak vulnerability.
- Access revoked globally for all foreign nationals; Anthropic complied immediately.
Why It Matters
Shows AI models can be abruptly cut off by government orders, disrupting enterprise deployments and raising compliance risks.