The news dropped Friday afternoon — and it’s a genuine relief: the US Commerce Department has lifted the export ban on Claude Mythos 5. Anthropic’s most powerful AI model can once again be deployed to over 100 US institutions.
What happened
Commerce Secretary Howard Lutnick wrote to Anthropic stating that he had determined ‘appropriate safeguards are in place to permit certain trusted partners to access the Claude Mythos 5 Model.’ Diplomatic language aside, the message is clear: Mythos 5 is no longer under quarantine.
Quick recap: on June 12, the Trump administration ordered Anthropic to shut off all access to Fable 5 and Mythos 5, citing national security concerns. The trigger was fears that the models could be exploited for cyberattacks — after Amazon and other partners flagged potential jailbreaks.
What Mythos 5 can do now
The clearance is limited: Mythos 5 is initially being deployed to a select group of cyber defenders and infrastructure providers. Anthropic itself calls Mythos 5 its ‘strongest cybersecurity model.’ Over 100 US institutions — including major companies and government agencies — are getting access.
Fable 5 is next in line
The bad news: Lutnick’s letter does not include permission for Fable 5, the powerful consumer-facing version of Mythos. Fable 5 has been offline for 15 days — and users have felt every one of them.
The good news: according to Axios, the Trump administration is close to lifting restrictions on Fable 5 as well. Insiders expect clearance could come as early as next week.
Why this matters
The Mythos clearance is more than a model update — it marks a real de-escalation in the standoff between the Trump administration and Anthropic. For two weeks, it was unclear whether the US government would permanently block Anthropic’s frontier models. Finding a resolution is a positive signal for the entire AI industry.
For the many users who’ve been without Fable 5 since June 12, one question remains: will the model return at full capability — or will those ‘appropriate safeguards’ come with noticeable limitations?
We’ll find out soon enough.
Sources: