Anthropic restores Claude Fable 5 and Mythos 5 after Trump administration lifts U.S. export controls

Less than three weeks after Washington forced Anthropic to pull two of its newest AI models from much of the market, the company is bringing them back. The U.S. Department of Commerce has lifted the export controls that halted access to Claude Fable 5 and Mythos 5, clearing the way for Anthropic to resume global availability and ending one of the most closely watched clashes between an AI company and the Trump administration.

“We’ve received notice that the Department of Commerce has lifted export controls on Claude Fable 5 and Mythos 5. We’ll begin restoring access tomorrow and will share an update soon. We’re grateful to our users for their patience, and to everyone who worked with us on redeploying the models,” Anthropic wrote Tuesday in a post on X.

The timing is significant. As prices for leading U.S. frontier models continue to climb, some Western companies have quietly started shifting AI workloads to lower-cost Chinese open source models. The temporary withdrawal of Anthropic’s newest models had fueled concerns that U.S. export restrictions were slowing American AI companies just as Chinese rivals were gaining momentum.

The reversal ends an extraordinary directive issued in mid-June that forced Anthropic to suspend access to Fable 5 and Mythos 5 for any foreign national, whether located inside or outside the United States. The restriction reached beyond customers and applied to Anthropic employees who were not U.S. citizens, underscoring how aggressively the government was applying national security rules to frontier AI systems.

Starting Wednesday, Fable 5 will once again be available worldwide through Claude, Claude.ai, and Claude Code. Anthropic said users on Pro, Max, Team, and selected Enterprise plans will receive access worth up to 50% of their weekly usage limits through July 7, CNBC reported.

The company said it has already restored access to Mythos 5 for selected U.S. organizations after receiving government approval on June 26. Anthropic plans to continue working with federal officials to extend availability through its Glasswing cybersecurity program, which gives vetted organizations access to advanced AI models for defensive security testing.

Anthropic said it is working to restore Fable 5 on Amazon Web Services, Google Cloud, and Microsoft Foundry as quickly as possible.

The policy reversal closes a tense chapter that raised broader questions across the AI industry. The restrictions arrived just as Chinese open-source AI models were gaining traction among developers and enterprises seeking lower-cost alternatives to expensive frontier models from U.S. labs. Some investors and technology executives warned that delaying American model releases risked giving Chinese competitors valuable time to narrow the performance gap.

The Commerce Department had already signaled that restrictions might ease. Last week, Commerce Secretary Howard Lutnick authorized Anthropic to release Mythos 5 to a limited group of companies and federal agencies after determining that sufficient safeguards were in place.

“Over the past two weeks, we have worked closely with Anthropic to analyze and approve Fable 5 to ensure alignment across the US Government and strengthen America’s leadership in AI,” Lutnick wrote Tuesday on X.

How Anthropic’s Claude Fable 5 launch turned into a showdown with Washington

The standoff began shortly after Anthropic unveiled Fable 5 and Mythos 5, describing them as its most capable models across several industry benchmarks. Fable 5 marked the first time the company released one of its highest-performing systems for broad public use, making the government order particularly disruptive for customers and developers.

Behind the scenes, Anthropic entered urgent discussions with officials in Washington after receiving the June 12 export directive. The company disclosed few details publicly during the negotiations. Reports later indicated that Anthropic co-founder Tom Brown took over discussions with the administration, replacing CEO Dario Amodei. Lutnick’s approval letter was addressed to Brown rather than Amodei.

Amodei has frequently clashed with the Trump administration over AI policy. He has publicly argued for stronger AI safety measures and supported former Vice President Kamala Harris during the 2024 presidential campaign, positions that reportedly strained relations with the White House.

The episode has exposed growing uncertainty around how frontier AI models will be regulated in the United States. AI companies say they are still trying to understand how government reviews will work and which officials are shaping national AI policy, particularly after David Sacks stepped down earlier this year as the administration’s crypto and AI czar.

President Donald Trump signed an executive order on AI in early June that called on developers to voluntarily submit advanced models for government evaluation before broad public release. Federal agencies were given 60 days to establish review procedures, leaving companies launching new models in a temporary policy gap.

OpenAI encountered many of the same issues. Last week, the company introduced three new AI models and said it was limiting initial access to “a small group of trusted partners” after consulting with the government before launch.

“We don’t believe this kind of government access process should become the long-term default,” OpenAI said at the time. “It keeps the best tools from users, developers, enterprises, cyber defenders, and global partners who need them.”

Anthropic’s restored rollout marks one of the first major signs that Washington is willing to relax restrictions once security concerns are addressed. It also removes a significant obstacle for developers and enterprise customers who had been waiting for access to the company’s newest AI models, at a time when global competition in frontier AI continues to intensify.