Trump Lifts Export Controls on Anthropic Claude

Anthropic's Claude Fable 5 and Mythos 5 are back after an 18 day export ban, but the fix reveals how fragile AI safety rules…

Anthropic's Claude Fable 5 and Mythos 5 models are back online after the U.S. government lifted export controls that had blocked access worldwide for 18 days, a reversal the company announced Tuesday that restores its flagship AI tools just as it pursues a confidential IPO filing.

What Changes Starting Wednesday

Users everywhere can once again reach Fable 5 through Claude.ai, Claude Code, Claude Cowork, and the broader Claude platform. Anthropic is sweetening the return: through July 7, subscribers on Pro, Max, Team, and select Enterprise plans get Fable 5 access covering up to half their weekly usage allotment before credits kick in. The company says it also plans to restore Fable 5 on Amazon Web Services, Google Cloud, and Microsoft Foundry, though it hasn't given a date for that rollout.

Mythos 5 access is more limited for now. A small set of U.S. organizations regained entry after the government approved it on June 26. Broader access, including for international partners under the Glasswing program, is still being negotiated with federal officials.

Why the Models Went Dark

The trouble traces back to a June 12 directive triggered by Amazon researchers, who documented a method for coaxing Fable 5 into producing dangerous outputs tied to discovering software vulnerabilities. Because the resulting prohibition applied to all foreign nationals, a category that swept in Anthropic's own researchers, the company decided its only option was to pull both models offline entirely rather than risk noncompliance.

Anthropic pushed back on the idea that Fable 5 was uniquely risky, saying its internal testing showed other widely used models, including OpenAI's GPT-5.5, could generate the same flagged outputs when prompted the same way.

Employees exit a federal government building in Washington at dusk.

The Fix That Satisfied Regulators

In response, Anthropic built a new safety classifier designed to catch the specific bypass technique Amazon's researchers had identified. In the company's own testing, it stopped that technique more than 99 percent of the time. The Commerce Department's Center for AI Standards and Innovation reviewed both the original safeguards and the new classifier and confirmed they worked, according to Anthropic. Any request to Fable 5 that still gets blocked will now be rerouted automatically to Claude Opus 4.8.

Officials Weigh In

Commerce Secretary Howard Lutnick said on X that his department spent two weeks reviewing Fable 5 with Anthropic to get the federal government aligned and to strengthen America's standing in the technology race. White House chief of staff Susie Wiles echoed that framing on X, telling the Wall Street Journal that the shared goal is deploying strong technology as fast and safely as possible.

A Rocky Moment for a Company Eyeing a Massive IPO

The shutdown landed at an awkward time for Anthropic, which has filed a confidential IPO prospectus disclosing a $47 billion revenue run rate and a $965 billion valuation. CNBC reported that co-founder Tom Brown took over as lead negotiator with the government in place of CEO Dario Amodei, a shift visible in Lutnick's correspondence, which was addressed directly to Brown rather than Amodei.

Will Mythos 5 Access Widen Further?

Fable 5 is back for everyone, but Mythos 5 remains restricted to a limited slate of domestic organizations. Whether the Glasswing program expands to more U.S. partners and international users depends on how ongoing talks with federal regulators play out, and Anthropic hasn't offered a timeline for that decision.