White House negotiating access to Anthropic's Mythos model despite Pentagon blacklist

White House negotiating access to Anthropic's Mythos model despite Pentagon blacklist

The White House and Anthropic are in active discussions to deploy the AI company's Mythos Preview model within federal agencies, despite ongoing Pentagon efforts to blacklist Anthropic as a supply chain risk, according to sources familiar with the negotiations.

Agencies may receive access to Mythos within weeks, two sources told Axios. The Office of Management and Budget confirmed it is examining the matter after receiving queries from agencies about whether they can use the model.

The Pentagon conflict

Anthropic remains in litigation with the Pentagon, which declared the company a "supply chain risk" and ordered defense contractors to remove its software from military workflows. The company is barred from Pentagon contracts but can conduct business with other federal agencies while the case proceeds.

The dispute centers on Anthropic's restrictions on model usage. The company prohibits its models from being used for mass surveillance or fully autonomous weapons development. The Pentagon argues these definitions are too vague and demands assurance it can deploy AI systems for "all lawful purposes."

"There's progress with the White House. There's not progress with [the Department of] War," one administration official said.

Why agencies want Mythos

Anthropic is rolling out Mythos Preview only to select organizations rather than the public, allowing them to assess what sources describe as "frightening cyber capabilities" and strengthen defenses.

Civilian agencies including the Departments of Energy and Treasury want access to identify vulnerabilities in critical infrastructure like the electric grid and financial system. "All the intel agencies use Anthropic. Every agency except War wants to," the first official said. "If you're the Department of Energy, you don't give a f*** about that. You're worried about the Chinese attacking the energy grid."

A second administration official accused Anthropic of using "fear tactics" with warnings about Mythos's hacking potential: "They're using this Mythos cyber weapon to find friendly ears in the government. They're succeeding."

Technical details

Pricing, context window size, and benchmark scores for Mythos Preview have not been publicly disclosed. The model remains in limited preview with no general availability timeline announced.

Both Anthropic and the Pentagon declined to comment.

What this means

The negotiations reveal a split in the Trump administration's approach to Anthropic. While some officials view the company as "woke doomsters" and supported the supply chain risk designation, even critics acknowledge Anthropic builds "best-in-class" tools for national security applications. One Defense official said talks continue only because "these guys are that good."

The outcome will test whether the government can compartmentalize vendor relationships—using a company's civilian tools while excluding it from military contracts—or whether Pentagon blacklisting effectively bars federal deployment regardless of technical merit.