Anthropic announced on Friday, June 26, 2026, that U.S. authorities had allowed access to its advanced cybersecurity model, Claude Mythos 5, to resume for a restricted group of organizations, two weeks after access had been completely suspended following an export-control directive issued by the U.S. government.

According to a letter sent to Anthropic by Commerce Secretary Howard Lutnick, certain government-approved organizations may once again use the model without requiring an export license. The letter was reviewed by several U.S. news organizations, including NBC News, CNBC, and Reuters. It states that the decision was based on the company’s “significant progress” in addressing the risks associated with the affected models, as well as on safeguards deemed adequate for access by trusted partners.

According to Reuters, CNBC, and NBC News, access will be granted to more than 100 organizations, including critical-infrastructure providers and institutions involved in cyber defense, many of which are part of Anthropic’s Project Glasswing program. The complete list has not been made public, and the authorities reserve the right to change the authorized entities at any time.

The initial restrictions were imposed on June 12, when the Department of Commerce ordered access to the Mythos 5 and Fable 5 models to be blocked for anyone who was not a U.S. citizen, regardless of whether they were located inside or outside the United States—including Anthropic’s own employees. The government cited the discovery of a specific technique that could allegedly bypass one of Fable 5’s protective safeguards, namely a mechanism designed to prevent the model from being used to identify software vulnerabilities. Authorities were also concerned that these capabilities could fall into the hands of intelligence services in countries such as China or Russia.

Because it was unable to distinguish in real time between American and foreign users, Anthropic decided to disable access to both models globally, describing the measure as the only way to comply with the directive. The company partly disputed the seriousness of the issue identified by the authorities, but continued negotiations with the government to restore controlled access.

Fable 5—the version intended for the general public, released only a few days before the ban—remains completely unavailable. Anthropic said discussions with the U.S. government are continuing, both on expanding access to Mythos 5 and on a possible broader relaunch of Fable 5.

Sources:

Reuters — U.S. Allows Anthropic to Release Mythos AI to “Trusted” U.S. Organizations
NBC News — U.S. Government Gives Anthropic Green Light for Limited Re-release of Mythos 5
CNBC — Trump Administration Allows Anthropic to Release Mythos AI Model to Some Companies and Government Agencies