Anthropic will soon start restoring access globally to its most powerful AI models, Fable 5 and Mythos 5, after the U.S. government elevated a restriction on wherein they may be launched, the company said Tuesday.
Over the past couple of weeks, the Trump administration has invoked national security worries to limit the ability of essential U.S. Tech companies to release advanced models, including those from Anthropic, which a few researchers feared couldbe exploited to bypass cybersecurity measures.
“We’ve attained notice that the Department of Commerce has lifted export controls on Claude Fable 5 and Mythos 5,” Anthropic posted on X. “We’ll start restoring access tomorrow.”
Just 4 days ago, the company stated it had obtained authorization from the government to permit a small group of American cybersecurity firms to access Mythos 5.
Secretary of Commerce Howard Lutnick stated in a June 26 letter to the corporation that “Anthropic has worked with the U.S. Government to deal with risks reated with the Covered Models,” Politico reported.
The government suddenly pressured Anthropic to cut off access to its cutting-edge artificial intelligence models on June 12 after founding vulnerabilities in the safeguards put in place to stop misuse of the tools.
On Tuesday, Lutnick informed Anthropic in a letter that the Trump administration had “withdrawn” its earlier restrictions on the introduce of the company’s models, Politico reported.
The letter showed that the Trump administration was satisfied, as a minimum for now, that Anthropic had “taken steps in close coordination with the U.S. Authorities to cope with the risks related to Claude Mythos 5 and Claude Fable 5.”
Like Anthropic, rival AI lab OpenAI has also agreed with Washington’s requests to restrict its very own launch of a new, powerful model referred to as GPT-5.6 to a limited set of approved partners.
“This isn’t quite the procedure that we suppose is most appropriate,” OpenAI CEO Sam Altman said Friday in a submit on X, along an explanation of the GPT-5.6 release.
Anthropic did now not straight away reply to a request for comment Tuesday.
Moreover, in a blog put up published Tuesday evening, Anthropic known as for the development of a standardized “framework” to both determine essential vulnerabilities in advanced models and respond to them.
The San Francisco-based AI lab will work with Amazon, Microsoft, Google and others on the effort.
“This issue will become more acute inside the coming months, as more models with effective cybersecurity (and other) capabilities are trained, assessed and introduced,” the blog submit stated.
New frontiers
The Trump administration published an executive order on June 2 demand the federal government to take numerous steps over the subsequent 2 months on AI and cybersecurity—such as creating a voluntary “framework” for private companies, such as Anthropic and OpenAI, to test and introduced their powerful “frontier” AI models in collaboration with the government.
Susie Wiles, the president’s chief of staff, posted Tuesday on X that the Trump administration was grateful for the cooperation from tech agencies, even though she didn’t call any.
“My gratitude to corporations across industries who persist to work carefully with the White House to enforce the President’s” executive order on AI and cybersecurity, Wiles stated. “This includes excellent work around advanced model access and guardrail testing and security.”
Earlier in the day, CIA Director John Ratcliffe evaluate the capabilities of the most advanced artificial intelligence models to nuclear weapons, in a tacit defense of the Trump administration’s recent tough line on controlling the release of the most powerful AI technology.
“In conversations with a few of the president’s other national security and economic security advisers, we are talking about the effect of those frontier AI models,” Ratcliffe stated during a speech at the AWS summit in Washington.
“It would be … not misplaced to refer to their capabilities as akin to digital nuclear weapons,” Ratcliffe stated.
