The U.S. Department of Commerce has officially removed export control restrictions on Anthropic’s Claude Fable 5 and Mythos 5 AI models. This decision concludes an 18-day period during which global access to these advanced systems was restricted.
Background on the Export Control Decision
On June 30, 2026, Commerce Secretary Howard Lutnick communicated in a letter to Anthropic’s Chief Compute Officer, Tom Brown, that a license is no longer necessary for the export, reexport, or in-country transfer of the Mythos and Fable models. This change follows initial actions by the Bureau of Industry and Security (BIS) on June 12, aimed at suspending access for foreign nationals due to concerns about national security and diversion risks.
The restrictions were initially prompted by warnings from Amazon executives and NSA reviewers who identified potential vulnerabilities in the models. Consequently, Anthropic had to disable both AI systems for all customers to comply with the directive, despite objections from over 100 cybersecurity experts who argued that the ban hindered essential security research.
Reassessment and Lifting of Controls
On June 26, a partial relaxation of restrictions was implemented, allowing limited access to Mythos 5 for selected government-approved organizations. This move was contingent upon the establishment of “appropriate safeguards.”
The final decision to lift export controls entirely was based on Anthropic’s commitments to detect and report security threats, collaborate with the government on release protocols, and disclose any malicious activities related to the AI models. The BIS reassessed the diversion risks and found them manageable, leading to the withdrawal of previous restrictions.
Resumption of AI Model Access and Future Outlook
Following the Commerce Department’s decision, Anthropic announced via X that customer access to Fable 5 would resume on July 1, with Mythos 5 access already partially restored the previous week.
Moreover, Anthropic took this opportunity to introduce Claude Sonnet 5, a new model offering a 1 million token context window and more cost-effective pricing tiers, continuing their innovation in AI architecture.
This development underscores the importance of balancing national security with technological advancement and highlights ongoing efforts to maintain secure yet open access to cutting-edge AI technologies.
