The US government has partially lifted restrictions on Anthropic’s Claude Mythos 5 AI model, allowing over 100 approved organizations to resume access following the implementation of new safety safeguards.

Key Points

• US Commerce Secretary Howard Lutnick authorized the redeployment of Mythos 5 to a select group of cyber defenders and infrastructure providers.
• The updated policy permits approved organizations and Anthropic to grant model access to foreign national employees.
• Restrictions remain in place for the consumer-facing Claude Fable 5 model while discussions with the White House continue.
• The initial June 12 directive was prompted by concerns regarding foreign access and potential security vulnerabilities identified by the National Security Agency and Amazon.
• Anthropic executives, including Tom Brown and Sarah Heck, negotiated the partial reinstatement during recent meetings with administration officials in Washington, DC.

Why it Matters

This development signals a shift toward a more rigorous, government-monitored approval process for the release of frontier artificial intelligence models. It highlights the growing tension between AI developers and federal regulators as the administration seeks to establish a formal policy framework for national security and export controls.