Trump eases – The Trump administration has loosened restrictions on Anthropic’s Claude Mythos 5, allowing access to more than 100 “trusted” U.S. organizations after a prior shutdown. The move stops short of a wider rollout and leaves uncertainty around Claude Fable 5.
By the time the letter landed, the practical question for Anthropic wasn’t whether its model could be used — it was who, exactly, would be allowed to use it.
In a letter sent to Anthropic cofounder and chief compute officer Tom Brown. US Commerce Secretary Howard Lutnick said the government would permit certain trusted partners to access Anthropic’s Claude Mythos 5. The change allows the company to grant access to more than 100 US organizations, including large corporations and government agencies. The letter was obtained by WIRED, and it had been reported earlier by Semafor.
Lutnick’s message was blunt about safeguards. He wrote that he had “determined that appropriate safeguards are in place,” adding: “Anthropic has worked with the U.S. government to address risks associated with the Covered Models. These efforts have yielded significant progress.”
The easing is also limited in a way that matters. The government stopped short of permitting a broader rollout of the model. In Lutnick’s letter, he pointed to the requirements outlined in the initial directive sent on June 12, saying those other conditions remain in effect.
What that means on the ground is visible in Anthropic’s response. Eduardo Maia Silva. an Anthropic spokesperson. said in a statement: “We received notice from the US government that Mythos 5. our strongest cybersecurity model. can be redeployed to a small group of cyber defenders and infrastructure providers.” He added that the company is working to provision the approved set of providers and “restore their access to Mythos 5 as quickly as possible.”.
Anthropic also signaled that its focus isn’t only on Mythos 5. Maia Silva said the company is “pleased to see this progress” and will keep working with the government to expand access to Mythos 5 while “make Fable 5 available for general use again.”
That second part is still unresolved. Lutnick did not say anything about the fate of Claude Fable 5. the consumer-facing version of Mythos that Anthropic released with significant additional safeguards. Anthropic is still in discussions with the White House about restoring access to Fable 5. and those talks are expected to continue over the weekend. according to a person familiar with the matter. Both parties are hopeful that resolving the dispute will help shape a lasting policy framework for future model releases.
The reversal comes only about two weeks after the White House sent an export control directive to Anthropic. That order required the company to limit foreign nationals from accessing Mythos and Fable 5 — including people working and living in the United States. In response, Anthropic disabled access to the models entirely.
The administration’s concern dates to how the rollout was moving before the shutdown. The Trump administration grew concerned after learning Anthropic granted access to a South Korean telecommunications firm it believed had ties to China. WIRED previously reported. Separately, Amazon and the National Security Agency also raised concerns to the White House that Fable 5 could be jailbroken. Officials say the combination of these events convinced them they needed to take action.
In recent weeks, Anthropic sent senior members from its cybersecurity and AI safety teams to Washington, DC to meet with Trump administration officials. Tom Brown and Sarah Heck, Anthropic’s public policy chief, led discussions with the US Department of Commerce.
The story’s immediate outcome looks like a partial reprieve: getting Mythos 5 back online is a step forward for both Anthropic and the White House. But the broader uncertainty — especially over how far the administration will go in controlling future model releases — is already part of the fallout. On Friday. OpenAI announced it was delaying the release of its upcoming GPT 5.6 models in response to a request from the Trump administration.
For Anthropic and the organizations now allowed in. the new permission comes with a familiar undertone: access is no longer just a business decision. It’s a negotiated boundary. defined by who is trusted. what safeguards are accepted. and which version of a model — Mythos 5 or Fable 5 — is considered safe to reach the next set of users.
Trump administration Anthropic Claude Mythos 5 Claude Fable 5 Howard Lutnick Tom Brown US Department of Commerce export control directive AI policy cybersecurity model
So they shut it down then “trusted” people get it again? Sounds like nothing changed just a different list.
I don’t even know what Mythos 5 is but if it’s cybersecurity then of course companies and agencies get access. Meanwhile the rest of us are still stuck with whatever crappy tools we have.
Wait so the government is basically deciding who gets to use the AI? That’s kinda scary, like who are the “trusted” organizations because I bet half of them are contractors that already had access anyway. Also what about Claude Fable 5… is that the one that got “uncertain” or am I mixing it up?
“Appropriate safeguards are in place” yeah sure. These letters always sound confident until something leaks. Plus they say it’s only a small group but also “more than 100” organizations which… is it small or not? I’m confused. And WIRED getting the letter means like, what, the rest of us just gotta assume it’s fine? lol