Two weeks after a sweeping export control order forced Anthropic to pull its most powerful cybersecurity models from the market, the Trump administration has restored Anthropic Mythos 5 access for more than 100 specific US government agencies and companies, including permission for non-American employees at those organisations to use the model.
Commerce Secretary Howard Lutnick sent a letter to Anthropic’s chief compute officer Tom Brown on Friday confirming the partial reversal. ‘I have determined that appropriate safeguards are in place to permit certain trusted partners to access the Claude Mythos 5 Model,’ Lutnick wrote, according to the letter seen by Semafor. The directive does not cover Fable 5, Anthropic’s other restricted cybersecurity model, which remains unavailable.
How the Ban Began, and Why It Hit All Customers
The original order arrived at 5:21 p.m. ET on 12 June, instructing Anthropic to suspend access to both Fable 5 and Mythos 5 ‘by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees,’ according to CNBC. To ensure compliance, Anthropic disabled both models for all its customers, not just foreign nationals, while confirming its other models would remain unaffected.
The ban caught Anthropic’s own non-American staff in its scope. Friday’s partial restoration explicitly extends to those employees as well, bringing them back inside the permitted circle alongside the more than 100 organisations on the approved list.
The directive invoked national security authorities, a framing that legal analysts at Greenberg Traurig described in a client alert as a novel regulatory challenge for companies deploying advanced frontier AI models. The episode has no direct precedent in how export control law has previously been applied to software.
Anthropic Mythos 5 Access Restored for Critical Infrastructure Operators
Lutnick’s letter cited ‘significant progress’ in intensive talks between the administration and Anthropic since the June 12 restrictions were imposed, according to Ynetnews, which reviewed the letter. Beyond the immediate partial lift, the letter sets out a forward-looking commitment: Anthropic has agreed to ‘work with the U.S. government on protocols and standards and releases’ for its models going forward, language reported by The Hill that signals an ongoing compliance relationship rather than a one-off concession.
Anthropic acknowledged the development in a post on X. ‘Since June 12, we’ve been working closely with the US government to restore access to Claude Mythos 5 and Fable 5,’ the company wrote. ‘Today, the government notified us that Mythos 5, our strongest cybersecurity model, can be redeployed to a set of US organizations that operate and defend critical infrastructure. We’re restoring access for these organizations quickly, and we’re continuing to work with the government to expand access to Mythos 5 and make Fable 5 available for general use again.’
A company spokesperson separately told The New York Times that Anthropic was working to restore access to the clients approved by the administration and was ‘pleased to see this progress.’
The background to the models’ removal matters here. A version of Mythos 5 was released publicly a couple of days before the ban, on the grounds that it contained stronger protections. Both Mythos 5 and Fable 5 were ultimately pulled after security researchers allegedly bypassed those guardrails with relative ease, giving the administration cause to act under its national security powers.
The timing of Lutnick’s decision is also worth noting in the wider context of AI regulation. His Friday letter arrived hours after Anthropic competitor OpenAI announced, at the federal government’s request, that it would release its GPT-5.6 family of models in phases rather than all at once. Two of the most capable AI labs in the country are now navigating government-imposed release schedules simultaneously, a development that would have seemed far-fetched only months ago.
For the organisations now cleared to use Mythos 5 again, the relief is real but partial. Fable 5 remains off the table, and Anthropic’s promise to restore it ‘for general use again’ comes with no timeline attached. The next indicator to watch: whether the administration moves to broaden the permitted list beyond the current 100-plus, or extends equivalent clearance to Fable 5.
