Anthropic's Fable 5 Hit by US Export Controls: Taken Down Globally Just Three Days After Launch

On June 12, 2026, the US Department of Commerce imposed export controls on Anthropic's latest AI models, Fable 5 and Mythos 5, citing national security, barring access by any foreign nationals, including Anthropic's foreign employees. Anthropic promptly took both models offline globally, just 72 hours after their June 9 release.

Timeline

• At 5:21 PM ET on June 12, Anthropic received a government directive to suspend foreign national access to Fable 5 and Mythos 5. The letter did not provide specific national security concerns.
• To comply, Anthropic disabled both models for all users. Other Claude models were unaffected.
• The government had previously privately urged Anthropic not to release the models; when that failed, it imposed controls.

Trigger: Amazon's 'Report' and Jailbreak Demo

• According to Axios, WSJ, and others, Amazon submitted a research report to the government claiming its researchers had successfully 'jailbroken' Fable 5, bypassing safety guards to obtain information usable for cyberattacks. Amazon CEO Andy Jassy spoke with Treasury Secretary Scott Bessent and other officials.
• At least five other companies also expressed similar concerns to the government.
• Anthropic countered that the demo revealed only a 'narrow, non-general' jailbreak, similar to vulnerabilities in other public models (e.g., GPT-5.5), and that such capabilities are routinely used by safety researchers.

Reactions

• Anthropic: CEO Dario Amodei called it a 'misunderstanding,' and the company is working to restore access. Anthropic began refunding affected Max subscribers, with a deadline of June 20.
• White House: Advisor David Sacks said export controls were imposed after Anthropic 'refused to fix the jailbreak vulnerability or take down the model.'
• Industry: LeCun criticized Amodei for spreading panic that led to foreign employees losing access. Security researchers mocked Anthropic for 'marketing itself as building a bomb, and the government officially declared it a bomb.'

Impact and Aftermath

• This marks the first time in AI history that a deployed commercial large model has been forcibly recalled by the government. Anthropic had planned an IPO in the fall (valued at ~$965 billion); the event caused its secondary market stock price to drop.
• Microsoft, due to Fable 5, required 30-day data retention and banned its own employees from using it, but offered it to customers (with manual opt-in).
• The incident highlights the export control risks facing cutting-edge AI models, boosting interest in open-source models. Zhipu AI announced the full release and open-sourcing of GLM-5.2 on the same day.
• Anthropic stated it disagrees with recalling a model over a 'narrow jailbreak,' arguing that such a standard would halt deployment of all frontier models.