Trump Administration Partially Restores Anthropic's Claude Mythos Access to US Organizations
Commerce Secretary Howard Lutnick permits Anthropic to grant Claude Mythos access to over 100 US organizations after the company disabled the model following an export control directive.
Last verified:
US Commerce Department Conditionally Lifts Claude Mythos Ban
The Trump administration has permitted Anthropic to restore access to Claude Mythos, its advanced cybersecurity-focused model, to more than 100 approved US organizations. According to Wired AI, Commerce Secretary Howard Lutnick issued a letter to Anthropic indicating he had “determined that appropriate safeguards are in place” for the deployment. The authorization marks a partial reversal of a June 12 export control directive that forced Anthropic to disable Claude Mythos and Claude Fable entirely after concerns about foreign national access.
Approved Access and Ongoing Negotiations
The newly authorized organizations include corporations and government agencies working in cyber defense and infrastructure protection. According to the Lutnick letter, Anthropic may now allow foreign national employees at approved organizations to access Claude Mythos—a permission previously prohibited. Anthropic spokesperson Eduardo Maia Silva told Wired AI the company is “working to provision the approved set of providers and restore their access to Mythos 5 as quickly as possible.”
The reinstatement covers only Claude Mythos. Claude Fable, the consumer-facing version released with additional safeguards, remains restricted. According to Wired AI, Anthropic is still in discussions with the White House over Fable’s status, with talks expected to continue.
Why This Matters
The partial lift signals the Trump administration’s willingness to calibrate restrictions rather than impose blanket bans—provided companies demonstrate sufficient risk controls. For Anthropic, restoration to 100+ organizations restores revenue from a critical customer segment while leaving the broader market access question unresolved. For federal agencies and defense contractors, the decision unblocks access to what Anthropic characterizes as its strongest cybersecurity capability. However, the continued restriction on Claude Fable suggests the government remains cautious about dual-use risks in consumer-grade deployments, even with safeguards. The ongoing negotiations over a “lasting policy framework,” as Wired AI notes sources described the parties’ intent, indicate both sides expect this precedent to shape future model release approvals.
Frequently Asked Questions
What models are affected by the export control directive?
Claude Mythos, Anthropic's advanced cybersecurity model, and Claude Fable, the consumer-facing derivative. Mythos has been partially restored; Fable's status remains under discussion with the White House.
Who can now access Claude Mythos?
Over 100 approved US organizations, including corporations and government agencies designated as 'cyber defenders and infrastructure providers.' Foreign nationals employed by these organizations may also access the model under the new framework.
Why did the US government restrict these models?
The White House grew concerned after learning Anthropic granted access to a South Korean telecommunications firm with alleged ties to China. The June 12 directive prohibited foreign national access, forcing Anthropic to disable both models entirely.