The US government has lifted export controls on Anthropic's most powerful artificial intelligence models, clearing the way for the company to restore global access to them, according to Anthropic and multiple outlets including the BBC, the Financial Times, The New York Times and Reuters.
The reversal covers Claude Fable 5 and Mythos 5, the models affected by the restrictions. Anthropic said the Department of Commerce lifted the controls, and the Financial Times described the move as the White House lifting a "ban" on the models. The Trump administration acted after what several sources, including SecurityWeek and the Pittsburgh Post-Gazette, characterized as a cybersecurity alarm.
Anthropic reacted warmly. "We're grateful," the company said, as reported by Benzinga, and it framed the return around what 9to5Google quoted as "extraordinarily strong" safeguards. Outlets including Firstpost and Let's Data Science reported that Anthropic overhauled or tightened its security safeguards as part of the restoration. According to Engadget, Anthropic planned to begin restoring users' access to Mythos and Fable on July 1.
The comeback is not entirely free of friction for customers. TweakTown and Android Headlines reported that users are getting fewer free days than Anthropic had originally offered, framing the restored access as coming "at a price."
The news landed alongside other Anthropic moves. AI Business reported the company launched Sonnet 5, while NAI500 reported the debut of Claude Science, an entry into life sciences that it said sent AI drug-discovery stocks tumbling. Coverage from ERP Today framed the episode as a test of new rules governing AI access.
Why it matters: the decision shows how directly government export policy now shapes who can use the most capable AI systems, turning access to frontier models into a fast-moving question of national security and commercial strategy.