Trump administration lifts Mythos 5 export ban for more than 100 US companies and agencies

The Trump administration partially lifted its export control ban on Anthropic's Claude Mythos 5 model on June 26, 2026, authorizing more than 100 U.S. companies and government agencies to access the powerful AI system. Commerce Secretary Howard Lutnick made the decision after determining that appropriate safeguards are in place to permit certain trusted partners to access the model—reversing a two-week-old emergency restriction that had pulled Mythos 5 from the market entirely.

What's new

Lutnick sent a letter to Anthropic's chief compute officer Tom Brown formally authorizing the release. Key terms of the decision:

• More than 100 U.S. government agencies and private companies cleared for access
• Non-American employees at approved U.S. organizations may now access Mythos 5
• Anthropic's own international staff can now use the model
• Fable 5, Anthropic's other suspended model, remains unavailable for general release

Anthropics stated it is "continuing to work with the government to expand access to Mythos 5 and make Fable 5 available for general use again," but provided no timeline for Fable 5's return.

Context

On June 12, the Trump administration imposed emergency export controls citing national security concerns. Security researchers had reportedly demonstrated a method to "jailbreak" Fable 5—bypassing its safety guardrails—and officials were concerned about model access reaching partners with Chinese connections, including a South Korean telecommunications provider. The directive suspended access to both Fable 5 and Mythos 5 for all users, including Anthropic's own international staff.

Anthropics responded with a public statement challenging the decision: the company argued it "disagree[d] that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people." Anthropic also warned that applying such a standard industrywide "would essentially halt all new model deployments for all frontier model providers." Despite its objections, the company complied with the directive.

The June 26 partial lifting reversed the ban's most sweeping provisions. Notably, it restores access for foreign national employees at approved U.S. organizations—one of the original ban's stricter elements.

Why it matters

The Mythos 5 episode is the first time the U.S. government has imposed and then selectively lifted export controls on a commercially deployed frontier AI model. It demonstrates that Washington now treats access to advanced AI as a strategic policy lever—comparable to semiconductor export controls—subject to national security review and conditional authorization.

The calibrated approach is telling: Mythos 5, designed for controlled enterprise and government deployment, received clearance. Fable 5, the more broadly available model, remains suspended pending further review. This distinction suggests regulators are beginning to differentiate frontier models by deployment context, not just capability.

Anthropics commitment to ongoing cooperation with the government on "protocols and standards and releases" points toward a regularized review process for future frontier model launches. For the AI industry broadly, the case sets a precedent: frontier models can be pulled from market and selectively re-released based on government security assessments, a dynamic that has no clear prior in software or consumer technology.