The National Security Agency is said to be using Mythos Preview, Anthropic’s recently announced model that it withheld from public release, even as the company’s tools sit inside a larger fight over who gets access to powerful systems and for what kind of control. The NSA appears to be among the undisclosed recipients of a model Anthropic limited to around 40 organizations, after Anthropic said Mythos was too capable of offensive cyberattacks to be released publicly.
Who Gets the Tools
Anthropic announced Mythos earlier this month as a frontier model designed for cybersecurity tasks, but said the model was too capable of offensive cyberattacks to be released publicly. Instead of broad access, the company limited it to around 40 organizations, and has publicly named only a dozen. The NSA appears to be among the undisclosed recipients and is said to be using Mythos primarily for scanning environments for exploitable vulnerabilities. The U.K.’s AI Security Institute has also confirmed it has access to Mythos.
That arrangement leaves ordinary people far from the decision-making while state agencies and select institutions get privileged access to tools built behind closed doors. The model is not being released to the public, but it is being used by agencies with the power to monitor, probe, and secure systems at scale.
The State’s Double Game
The news comes weeks after the NSA’s parent agency, the Department of Defense, labeled Anthropic a “supply-chain risk,” after the company refused to allow Pentagon officials unrestricted access to its model’s full capabilities. The Pentagon’s dispute originated when Anthropic refused to make Claude available for mass domestic surveillance and autonomous weapons development.
At the same time, the U.S. military’s expanding use of Anthropic’s tools comes as it simultaneously argues in court that those tools can threaten national security. That is the familiar machinery of power talking out of both sides of its mouth: one arm of the state warns of danger while another seeks access.
The dispute shows how the apparatus tries to secure control over emerging technology, not for public benefit, but for surveillance and weapons development. Anthropic’s refusal to hand over unrestricted access did not end the relationship; it merely shifted the terms of the struggle over who gets to use the machine and how far the state can reach.
Power Reaches for the Model
The NSA’s access to Mythos comes as Anthropic’s relationship with the Trump administration appears to be thawing. Last Friday, Anthropic chief executive Dario Amodei met with White House chief of staff Susie Wiles and Secretary of the Treasury Scott Bessent. The White House reportedly called the meeting productive.
That meeting sits alongside the broader pattern: corporate executives and top officials negotiating access, risk, and leverage while the public gets none of the control. The company that withheld Mythos from public release is still making it available to a small circle of institutions, including a spy agency and the U.K.’s AI Security Institute.
TechCrunch has reached out to the NSA for comment, and Anthropic declined to comment. The silence fits the arrangement. The people most affected by these systems are not the ones invited into the room when access is decided, and the institutions with the most power are the ones already inside it.