After two weeks of tense negotiations with the Trump administration, Anthropic’s Mythos 5 is back in action, at least for a select group of organizations. Fable 5, the public-facing Mythos-class model, remains stuck in regulatory limbo with no timeline for release.
The breakthrough came in a letter from Commerce Secretary Howard Lutnick to Anthropic co-founder Tom Brown, dated June 26, which The Verge reviewed. Lutnick wrote that a “revision to the license requirements” was justified because Anthropic had “worked with the U.S. government to address risks” associated with both Mythos 5 and Fable 5.
Anthropic spokesperson Danielle Ghiglieri confirmed the development: “We have received notice from the US government that Mythos 5, our strongest cybersecurity model, can be redeployed to a small group of cyber defenders and infrastructure providers. We are working to provision the approved set of providers and restore their access as quickly as possible.”
What changed, and what didn’t.
The US government did not lift the export control directive it imposed on Anthropic two weeks ago, which barred any foreign national, including Anthropic’s own employees, from accessing either model. Instead, it carved out an exception: a select group of organizations approved for Mythos 5 access, mirroring the same limited preview model the administration granted OpenAI for GPT-5.6 earlier this week.
Under the exception, both US and non-US nationals working at approved organizations, and Anthropic employees who are not US nationals, can access Mythos 5. The export ban on Fable 5 remains fully in place.
“These efforts have yielded significant progress,” Lutnick wrote. “In addition, Anthropic has committed to work with the U.S. government on protocols and standards and releases for Mythos-class models. In light of this progress … I have determined that appropriate safeguards are in place to permit certain trusted partners to access the Claude Mythos 5 Model.”
Pressure from all sides.
The administration faced mounting pressure to loosen its stance. Competitors’ cybersecurity models were pulling ahead of Mythos 5 on key benchmarks. Top US government departments, including the National Security Agency, had lost access to a model they depended on. And industry fears were growing that China could capitalize on the sidelining of America’s leading AI labs.
Now Anthropic has essentially the same deal OpenAI secured: a limited preview for approved enterprises and government agencies, with general availability deferred to the administration’s discretion. Whether Fable 5, the model the public would actually use, ever gets released remains an open question.
Sources: Anthropic’s Mythos 5 is back (The Verge, June 2026)