The Trump administration has ended export controls on Anthropic's AI models, resolving a dispute that left the $1 trillion company unable to distribute its latest technology internationally. The Department of Commerce notified Anthropic on Tuesday evening that restrictions on Mythos and Fable models would be lifted, allowing the San Francisco-based firm to re-release Fable 5 to the general public. The ban, imposed on June 12, followed discovery of a security vulnerability that could bypass the model's safety protections.

Commerce Secretary Howard Lutnick signed off after Anthropic agreed to proactively detect and address security risks. The restrictions had inflamed tensions between Silicon Valley and the White House, with tech executives criticizing the administration's unpredictable approach to AI regulation. Foreign governments also complained that the ban unfairly punished allies. The episode highlighted broader concerns about how frontier AI models might be exploited for cybersecurity threats before their risks are fully understood.

The resolution allows Anthropic to resume global distribution of its advanced models while maintaining tighter controls for enterprise customers. The company had already released Mythos 5 to about 100 pre-vetted partners under the original framework. Open AI has similarly faced restrictions on its new GPT-5.6 model, which remains limited to government-approved partners. Both companies are navigating an evolving regulatory landscape as the administration balances innovation with national security concerns.

The lifting of restrictions signals a pragmatic compromise between AI developers and regulators. Rather than maintaining a hard line, the administration opted for a negotiated solution that preserves some government oversight while allowing commercial development to proceed.