Anthropic Restores Mythos 5 Access for Limited Group of Organizations
Anthropic has received U.S. government approval to redeploy its strongest cybersecurity model, Mythos 5, but only to a select group of organizations, while the public-facing Fable 5 remains on hold.
After two weeks of intense negotiations with the Trump administration, Anthropic has partially regained access to its most powerful cybersecurity model, Mythos 5. According to a letter obtained by The Verge, Commerce Secretary Howard Lutnick announced a revision of license requirements following Anthropic's cooperation with the government to address risks.
Anthropic spokesperson Danielle Ghiglieri confirmed that Mythos 5 can now be redeployed to a small group of cyber defenders and infrastructure providers. The export control directive, which barred foreign nationals (including Anthropic employees) from accessing the model, has not been lifted. Instead, the government made an exception, similar to its recent handling of OpenAI's GPT-5.6.
Pressure on the administration grew as competitors' cybersecurity models advanced and U.S. agencies like the National Security Agency lost access to Mythos 5. Concerns also centered on potential Chinese AI progress while top U.S. labs were sidelined.
Anthropic now has the same deal as OpenAI: a limited preview for approved organizations. Both companies hope for broader availability soon, but the final decision rests with the Trump administration. OpenAI stated in a blog post that this approach isn't a long-term solution but a short-term step toward wider access.
Lutnick's letter emphasized that all other requirements from the June 12 letter remain in effect, and he reserves the right to adjust the scope of license requirements if circumstances change.
