US Government Partially Lifts Restrictions on Anthropic's Mythos 5 AI Model

The U.S. government has revised its license requirements for Anthropic's Mythos 5 AI model, allowing over 100 specific U.S. companies and government agencies limited access. The move follows a two-week ban on the cybersecurity-focused model, though restrictions on the public-facing Fable 5 model remain in place.
US Government Partially Lifts Restrictions on Anthropic's Mythos 5 AI Model

US Government Partially Lifts Restrictions on Anthropic’s Mythos 5 AI Model The U.S. government has partially reversed an abrupt clampdown on Anthropic’s most advanced cybersecurity AI, Mythos 5, reopening access for a tightly controlled group of organizations while keeping broader public use on hold.

The June 12 shutdown

On June 12, the Commerce Department ordered export-style restrictions on Anthropic’s Mythos 5 and its public-facing sibling Fable 5, prompting the company to pull both models from the market and stunning the AI industry and U.S. allies that relied on their cyber capabilities. The move was framed as a response to diversion and security risks after researchers allegedly bypassed model guardrails.

Two weeks of negotiation

Over the following two weeks, Anthropic entered intensive talks with the Trump administration to address national security concerns and redesign access conditions for what officials describe as the firm’s “strongest cybersecurity model.” According to a letter reviewed by multiple outlets, Commerce Secretary Howard Lutnick said the company had “worked with the U.S. government to address risks” associated with Mythos 5 and Fable 5 and that these efforts had “yielded significant progress.”

June 26: Limited reopening for Mythos 5

In a June 26 letter to Anthropic co‑founder Tom Brown, Lutnick announced a revision to license requirements, allowing Mythos 5 to be exported or accessed by entities listed in an annex and their foreign-national employees, as well as Anthropic’s own non‑U.S. staff. “Accordingly, a license will no longer be required” for those approved entities, the letter states, while warning that Lutnick “reserve[s] the right to reevaluate and adjust the scope” and even change the access list “at any time.”

Anthropic confirmed it had “received notice from the US government that Mythos 5 … can be redeployed to a small group of cyber defenders and infrastructure providers” and said it was “working to provision the approved set of providers and restore their access … as quickly as possible.”

What remains restricted

Fable 5, the consumer-oriented version of Mythos 5, is still offline, with no agreed timeline for general release. Export controls and license requirements remain in place for all non‑approved organizations worldwide, leaving most of the AI ecosystem—and many foreign partners—locked out of Anthropic’s most powerful tools for now.

Continue reading https://foxvector.com/stories/019f078a-1586-12af-72d2-1fdf6f6fef03

Write a comment