Key Points
- White House negotiating Anthropic’s Mythos Preview access for federal agencies
- Anthropic remains blacklisted by Pentagon as supply chain risk amid litigation
- Government agencies may gain Mythos access within weeks despite restrictions
The White House is negotiating with Anthropic to give federal agencies access to the company’s powerful new artificial intelligence model, Mythos Preview, even as the Pentagon maintains a blacklist against the firm, according to sources familiar with the discussions.
The talks come at a time when Anthropic faces ongoing litigation with the Department of Defense, which has declared the company a supply chain risk. Despite this, multiple government agencies have expressed interest in evaluating Mythos Preview, a model the company says has significant cyber capabilities that could either strengthen national security or pose serious risks if misused.
EVENT
Infosec Reimagined
Infosec Reimagined 2026 is the premier information security summit where top leaders—CISOs, CROs, CIOs, CTOs and risk executives—converge to redefine cyber resilience.
EVENT
CIO Prism
CIO Prism unites forward-thinking technology leaders to exchange transformative insights, shape digital strategies, and foster innovation, empowering enterprises to excel in an era of rapid technological change.
According to multiple reports, that discussions are progressing and federal agencies may gain access to Mythos in the coming weeks.
What Mythos Preview offers
Mythos Preview is Anthropic’s latest AI model, designed with advanced capabilities that the company itself describes as potentially dangerous in the cyber domain. Unlike the firm’s publicly available Claude chatbot, Mythos is being rolled out only to a select group of companies and organisations.
The restricted release allows these partners to assess the model’s capabilities for offensive and defensive cyber operations and strengthen their security systems accordingly. Several government agencies have now asked whether they can join this limited access programme.
In response to these queries, the Office of Management and Budget sent an email to agencies, first reported by Bloomberg, stating that the matter was under review.
Pentagon dispute complicates access
The negotiations are complicated by Anthropic’s strained relationship with the Pentagon. The Department of War designated Anthropic a supply chain risk and ordered companies working with the military to remove its software from their systems.
Anthropic is currently barred from Pentagon contracts but remains eligible to work with other parts of the federal government while the legal dispute continues.
Advertisement
Axios cited one administration official, speaking on condition of anonymity, who said progress was being made with the White House but not with the Department of War.
A second administration official said the government had a responsibility to evaluate every model to understand where the frontier of technology stood.
Security versus access
The tension shows a broader debate within the administration about how to handle AI companies whose tools may be both essential and risky. Even officials critical of Anthropic acknowledge privately that the company is building technology that could significantly aid national security, but also cause harm if it falls into the wrong hands.
Anthropic has not publicly commented on the discussions. The company’s decision to limit Mythos Preview to vetted partners rather than releasing it publicly suggests internal concerns about misuse.
The outcome of these negotiations could set a precedent for how the federal government engages with frontier AI companies, particularly those facing regulatory or legal scrutiny from one part of the government while seeking contracts with another.
Your Questions, Answered
What is Anthropic’s Mythos Preview?
Mythos Preview is Anthropic’s latest AI model with advanced cyber capabilities. Unlike the company’s public Claude chatbot, it is available only to select organisations that can assess its potential for offensive and defensive cyber operations.
Why is Anthropic blacklisted by the Pentagon?
The Department of Defense designated Anthropic a supply chain risk and ordered military contractors to remove its software. The company is in ongoing litigation over this designation but can still work with non-military federal agencies.
Can federal agencies use Anthropic products?
Currently, Anthropic is barred from Pentagon contracts but can work with other parts of the federal government. The White House is negotiating terms under which agencies may access Mythos Preview.
When will federal agencies get access to Mythos?
According to sources familiar with the discussions, federal agencies may gain access to Mythos Preview within the coming weeks, pending the outcome of ongoing negotiations.