Anthropic has regained global access to its most advanced AI models, Claude Fable 5 and Mythos 5, after the Trump administration lifted export controls. The move ends a two-week ban that had spooked the company and prompted safety testing negotiations.
The restrictions were reportedly triggered by White House concerns over national security risks posed by frontier AI systems. The reversal indicates a tentative truce between the AI startup and the U.S. government, with both sides agreeing to implement new safety protocols.
Anthropic confirmed the models will now be released globally, though details of the testing framework remain unclear. The company has not disclosed specific metrics or thresholds for the safety evaluations, according to reports.
The episode highlights ongoing friction between AI developers and regulators balancing innovation against security concerns. Other AI firms, including OpenAI and Google DeepMind, are likely watching closely as precedent is set.
Some critics argue the ban was an overreach that could have ceded global AI leadership to China. Others contend the administration should have maintained tighter restrictions given the models' capabilities.