The NSA is using Anthropic's most powerful AI model Mythos

The NSA is reportedly using Anthropic’s top model, Mythos Preview.

That comes despite Pentagon concerns that Anthropic poses a security risk and should be blocked as a vendor.

Anthropic says access is limited under Project Glasswing, while broader government deployment is also being discussed with White House officials.

The report underscores rising government reliance on advanced AI alongside sharper security and oversight concerns.