The federal government is quickly changing its tune on advanced artificial intelligence regulations. Just two weeks after forcing Anthropic to pull its most advanced cybersecurity models from the public market, the Trump administration decided to soften its strict position. Officials are now greenlighting a specialized plan that lets Anthropic distribute its powerful Claude Mythos 5 model to a select group of critical organizations.

According to breaking reports from Semafor and Reuters, the new directive grants access to more than 100 specific United States government agencies and trusted private companies. In a major policy reversal, the administration is even allowing non-American employees at these approved organizations to use the software. This detail marks a massive change from the original ban, which strictly barred any foreign nationals from touching or interacting with the security models over extreme espionage fears.

The policy shift became official when Commerce Secretary Howard Lutnick sent a formal letter to Anthropic Chief Compute Officer Tom Brown. In the message, Lutnick stated that the government verified that appropriate safety safeguards are finally in place. Because of those verified protections, the administration decided to permit these trusted partners to log back into the platform.

Interestingly, the new government directive completely ignores Fable 5. Fable 5 is a specialized variant of the Mythos 5 framework that Anthropic released just days before the federal crackdown. The startup built Fable 5 with extra security layers designed to prevent digital criminals from weaponizing its code-scanning abilities. Despite those design intentions, Anthropic had to pull both models off the market when independent security researchers easily bypassed the guardrails during testing. The government has not yet addressed when or if Fable 5 will get a similar clearance.

The visual layout featured in image_be0543.jpg shows how Anthropic is managing this transition, but a closer look at the active software rollout can be seen in image_be0998.jpg. The phone screen in image_be0998.jpg shows the prominent Claude Mythos branding that sits at the center of this geopolitical tug-of-war.

Anthropic publicly confirmed the policy breakthrough in an official social media post on X. The company shared that its engineering teams have worked around the clock with federal officials since June 12 to resolve the security standoff. The startup celebrated the notification from Washington, calling Mythos 5 its strongest cybersecurity model to date. They noted that the system will immediately go to work helping defenders protect and shield critical American infrastructure from incoming digital threats. Moving forward, Anthropic plans to continue its close collaboration with federal regulators to widen access to Mythos 5 and eventually return Fable 5 to general availability.