AI & Tech · Washington and San Francisco
US allows Anthropic to restore Mythos 5 for trusted partners, Fable 5 may be unblocked soon
The Commerce Department authorized Anthropic to redeploy its most powerful cybersecurity model to a limited set of US organizations on Friday, partially reversing a two-week national security block, while negotiations to release the public-facing Fable 5 continue.
AI text based on reports by Reuters, Axios, Deutsche Welle and other publishers. Changed . How it is made
For membersThe first 30 seconds for everyone. Synthetic voice (AI).
In brief
The Commerce Department letter allows over 100 companies and federal agencies to access the powerful cybersecurity model, restricted since June 12.
The public-facing version remains suspended, yet the administration may lift restrictions as early as next week, pending Pentagon and NSA approval.
OpenAI restricted access to its newest models to a vetted US group at Washington's request, calling it a temporary measure and opposing it as a permanent norm.
Co-founder Tom Brown led talks while Dario Amodei stepped back; Treasury Secretary Bessent and Commerce Secretary Lutnick helped de-escalate the four-month standoff.
- 12 JunUS Commerce Department orders Anthropic to restrict access to Mythos 5 and Fable 5 over national security risks.
- 13 JunAnthropic disables global access to both models to comply with the export control directive.
- 26 JunCommerce Secretary Lutnick authorizes Mythos 5 release to a limited set of trusted US organizations.
- 27 JunAxios reports the administration may lift Fable 5 restrictions within the coming week; talks continue.
Two-week export control freeze
On June 12, the Trump administration invoked export controls to force Anthropic to disable both Mythos 5 and Fable 5, its most advanced AI models, to all users. The government cited fears that security guardrails could be circumvented, potentially allowing foreign military intelligence to misuse the systems. Anthropic, unable to distinguish domestic from foreign users in real time, shut off global access on June 13. The abrupt withdrawal of a top-tier model already in users' hands drew heavy criticism and reignited debate over US technology leverage.
Partial restoration for Mythos 5
Commerce Secretary Howard Lutnick informed Anthropic on June 26 that Mythos 5 could be released to "certain trusted partners" after the company had worked with the government to address risks. An Anthropic statement said the model would be redeployed to "a small group of cyber defenders and infrastructure providers" who operate critical US systems. A source told Reuters that more than 100 companies and institutions, including many Fortune 500 firms, will now have access. Foreign cybersecurity agencies in Europe and Asia remain excluded for now.
“Since the issuance of my June 12 letter, Anthropic has worked with the U.S. government to address risks associated with the Covered Models. These efforts have yielded significant progress.”
Fable 5's uncertain future
The letter did not alter the suspension of Fable 5, the public-facing version of the same underlying AI model. Axios reported on June 27, citing a source close to the situation, that the administration could lift limits on Fable 5 as soon as this coming week, though the Pentagon and National Security Agency have yet to sign off. Anthropic expects to restore general access soon, and talks are scheduled to continue through the weekend. Fable 5 had earned rapid acclaim for deep reasoning and coding before it was pulled just three days after launch.
Government tightens grip on frontier AI
OpenAI also disclosed that the US government demanded restricted access to its new GPT-5.6 models, allowing only a vetted small group of partners in the US. The company said it is working to extend access to other countries next week but expressed opposition to such controls becoming the norm. OpenAI CEO Sam Altman wrote that safety testing is fine but added a clear reservation.
“I just don't like the idea of the government picking the customers.”
A Commerce Department spokesperson said the administration worked diligently to ensure the US remains the global leader in AI while safeguarding security. The developments mark a new phase in Washington's direct oversight of releases of frontier AI systems.
Behind the negotiations
Senior Anthropic executives, including co-founder Tom Brown, met with Lutnick and other officials in recent days, according to Bloomberg. CEO Dario Amodei took a hands-off role, a shift that people close to the talks said reduced friction, contrasting with his earlier direct engagement with the White House. Treasury Secretary Scott Bessent was also credited with helping defuse the tension. Defense Secretary Pete Hegseth had previously labeled Anthropic a "supply-chain risk to national security" after clashing with Amodei over Pentagon use of Claude.
Ask about this text
Sources and context
- Howard Lutnick
- US Commerce Secretary who signed the letter partially restoring Mythos 5 access
- Dario Amodei
- CEO of Anthropic who took a hands-off role in the negotiations
- Tom Brown
- Anthropic co-founder who led the talks with the US administration
- Pete Hegseth
- US Defense Secretary who earlier labeled Anthropic a supply-chain risk
- Scott Bessent
- US Treasury Secretary who helped de-escalate the administration's standoff with Anthropic
- Sam Altman
- CEO of OpenAI who criticized the government's role in picking AI customers
Pollar is free and carries no ads in the text. Its members keep it going.
Support Pollar