Anthropic said Friday it had received authorization from the U.S. government to allow a small group of American cybersecurity firms to access its powerful AI model Mythos 5.
Access to Mythos 5 had been blocked by U.S. authorities over national security concerns.
A company spokesperson said Anthropic would continue discussions with the Trump administration to "expand access to Mythos 5 and make Fable 5 available again" to the general public.
The government abruptly forced Anthropic to cut off access to its two cutting-edge artificial intelligence models on June 12 after discovering vulnerabilities in safeguards designed to prevent misuse of the tools.
The move against Anthropic has drawn accusations of government overreach.
Commerce Secretary Howard Lutnick said in a letter to the company, cited by Politico and Reuters, that "Anthropic has worked with the U.S. government to address risks associated with the Covered Models."
"These efforts have yielded significant progress," he said.
Commerce Department spokesperson Benno Kass said officials had "worked diligently to ensure America remains the global leader in AI while safeguarding our security."
Anthropic angered President Donald Trump's team in an earlier clash by refusing to allow its technology to be used for mass surveillance and autonomous weapons, prompting the Pentagon to cancel its contracts with the company.
Anthropic's main rival, OpenAI, launched its new model, GPT-5.6, earlier Friday, with restricted access and validation handled by the government on a client-by-client basis.
OpenAI chief Sam Altman wrote on X that "this isn't quite the process that we think is optimal."
"I believe the government shares most of our goals, and that they are overall doing a good job in a very difficult situation," he said.
Under pressure over the novelty of advanced AI capabilities, Trump earlier this month signed an executive order setting up a voluntary federal review of national security risks in advanced AI models before their release.
The White House has said little about how it will enforce the executive order, which companies are understood to be participating in voluntarily, or which models would fall under the review rules.