Back to feed

Anthropic Confirms: US Government Partially Lifts Ban on Mythos 5

After the company fueled security concerns itself, the government under President Donald Trump issued an 'export ban.' Now, some companies can access the AI model again.

Anthropic Confirms: US Government Partially Lifts Ban on Mythos 5
  • Home
  • News
  • Software & Development

Add t3n as a preferred source on Google

Don't miss any news!

After the company had fueled security concerns itself, the government under President Donald Trump issued an "export ban." Now, some companies can access the AI model again.

Two weeks after Anthropic had to shut down its latest AI models due to an "export ban," the company is now allowed to provide Mythos 5 again, albeit in a limited capacity. According to Reuters, the US government has granted access to 100 selected US agencies and companies. Also, Anthropic employees outside the US, who were previously also affected by the ban, can now use the model again.

What Was the Real Reason for the Ban?

In early April, Anthropic introduced its new flagship model, Mythos, while simultaneously warning about its capabilities. In tests, it was said to be able to identify vulnerabilities in operating systems and browsers with exceptional precision. Instead of a broad release, Anthropic launched the controlled "Project Glasswing" and made the model available only to selected companies like Amazon, Apple, Google, and Microsoft so that they could optimize their cybersecurity. The recently released model Fable 5 was designed as a safer, freely accessible version that blocks responses in high-risk areas. Meanwhile, previous Mythos users gained access to Claude Mythos 5, a new version with fewer security restrictions but similar functionality.

The Trump administration initially justified the temporary "export ban" for both models with security concerns. However, as Wired reports, there were already objections regarding certain companies' access to the Mythos model. The trigger may have been the expansion of "Project Glasswing" in early June 2026. At that time, Anthropic had made the AI model accessible to around 150 additional companies in over 15 countries. According to Reuters, this included several South Korean corporations like Samsung, SK Hynix, and SK Telecom. The latter was particularly suspicious to the Trump administration due to its potential connections to China. Internal sources indicated that US authorities had instructed Anthropic to deny SK Telecom access to Mythos, although the company itself denied having any ties to China.

Mythos 5 Limited Availability Restored

Now the US government has relented. "I have concluded that appropriate security measures are in place to allow certain trusted partners access to the Claude Mythos 5 model," Commerce Secretary Howard Lutnick reportedly wrote in a statement to Tom Brown, the Chief Compute Officer of Anthropic. However, the government did not address the release of Fable 5.

Anthropic confirmed the decision in a post on X: "Since June 12, we have been working closely with the US government to restore access to Claude Mythos 5 and Fable 5. Today, the government informed us that Mythos 5, our most powerful cybersecurity model, can again be made available to a number of US organizations that operate and protect critical infrastructure," the post states. "We will restore access for these organizations promptly and continue to work with the government to expand access to Claude Mythos 5 and make Fable 5 available for general use again."