The U.S. government on Friday removed a block on Anthropic’s powerful Claude Mythos 5 model, permitting the company to distribute the model to more than 100 U.S. institutions. The decision was communicated in a letter sent to Anthropic on Friday afternoon.
Two weeks earlier the administration had imposed export controls on Mythos, prompting a shutdown of the model and its related Fable 5 after warnings from Amazon and other companies that the systems could be "jailbroken" for malicious uses. The recent letter does not address the status of Fable 5; that model is a weaker variant of Mythos and had briefly been the most powerful AI model widely available to consumers. Sources close to the discussions said parties are moving toward a release of Fable as well, but the timeline remains unclear.
In the letter, Commerce Secretary Howard Lutnick wrote to Anthropic’s chief compute officer Tom Brown that he had "determined that appropriate safeguards are in place to permit certain trusted partners to access the Claude Mythos 5 Model," citing "significant progress" in the intensive daily negotiations between the government and the company since the block was implemented. Lutnick also noted that Anthropic has committed to work with the U.S. government on protocols, standards and release procedures for its models.
On the same day, Anthropic’s chief competitor, OpenAI, released its latest model, GPT-5.6, to a short list of government-approved partners.
Under the new arrangement, a license will no longer be required to export, reexport, or transfer within the country (including deemed exports and reexports) the Claude Mythos 5 Model to entities listed in Annex A to the letter and their foreign national employees, or to Anthropic’s foreign national employees.
Why this matters
Lutnick’s letter signals the start of a nascent regulatory regime giving the U.S. government leverage over the distribution of frontier AI models. Leaders of frontier AI labs have warned that such oversight could slow them in the intense global AI race, but Commerce Department spokesman Benno Kass highlighted the speed with which the government addressed its security concerns about Anthropic: "In just two weeks, we have worked diligently to ensure America remains the global leader in AI while safeguarding our security," Kass said.
The oversight framework is still being built, and many potential users — from non-U.S. governments and firms to individual consumers — remain uncertain when they will gain access to Mythos or Fable. European officials and other U.S. allies have voiced frustration at becoming dependent on Washington’s decisions.
As initially reported by Semafor, U.S. officials had expressed concern that Mythos had been provided to partners with ties considered too close to China, reportedly referring to a South Korean telecommunications provider.
Next steps
The current authorization applies only to the more than 100 U.S.-based entities named in the letter’s Annex A. The fate of Fable 5 and access for foreign users remain unresolved, and detailed protocols, standards and longer-term regulatory mechanisms are still under development.
The episode illustrates how rapidly a national and international regulatory practice is being shaped to limit or permit access to strategically important AI technologies.
