White House to Grant Federal Agencies Access to Anthropic's Mythos AI Model

The White House is moving toward granting major federal agencies access to Anthropic's new artificial intelligence model Mythos, even as the company remains locked in litigation with the Pentagon over a national security designation, according to reporting by Bloomberg News and the Financial Times. Gregory Barbaccia, the federal chief information officer at the Office of Management and Budget, sent an email to cabinet department officials on Tuesday indicating that the OMB was establishing safeguards to allow agencies to begin using the model. The email, whose subject line read "Mythos Model Access," stopped short of confirming that access would be granted or specifying a timeline. Anthropic announced Mythos on April 7, 2026, and has restricted its release to a select group of organizations under an initiative called Project Glasswing, citing the model's ability to identify and exploit vulnerabilities "in every major operating system and every major web browser." Early access recipients include technology companies such as Nvidia, Microsoft, Google, and Apple, as well as financial institutions including JPMorgan Chase. The White House and Anthropic did not immediately respond to Reuters requests for comment.

Pentagon feud complicates government-wide access push The negotiations over Mythos access unfold against a backdrop of deep institutional conflict between Anthropic and the Department of Defense, which declared Anthropic a "supply chain risk" and barred it from defense contracts following a dispute over limits on the military's use of artificial intelligence. Anthropic sued to challenge that designation and won a preliminary injunction from a California court, allowing it to continue working with non-military government agencies while the case proceeds. A parallel case is ongoing in a federal appeals court. One administration official told Axios that there was "progress with the White House" but "not progress with the Department of War." A second official acknowledged the government "has a responsibility to evaluate every model to see where the frontier of tech is," while also accusing Anthropic of using "fear tactics" to find allies within the government. The same official said: „They're using this Mythos cyber weapon to find friendly ears in the government. They're succeeding.” — unnamed administration official via Axios Civilian agencies including the Departments of Energy and Treasury have been among the most eager to gain access, with officials arguing the model could help identify vulnerabilities in critical infrastructure such as the electric grid and financial system before adversaries exploit them.

UK institute confirms model's autonomous hacking capabilities Independent assessments of Mythos have lent credibility to some of Anthropic's more alarming claims about the model's capabilities, while also urging caution about overstating its real-world threat. Britain's AI Security Institute evaluated Mythos in controlled simulations and found it could "execute multi-stage attacks on vulnerable networks and discover and exploit vulnerabilities autonomously — tasks that would take human professionals days of work," according to a report cited by Bloomberg. The institute noted it was the first model to solve a lengthy hacking test it had devised, one that would take an experienced human approximately 20 hours to complete. However, the AISI concluded it "cannot say for sure" whether Mythos would be able to attack well-defended systems in real-world conditions. Jeetu Patel, president and chief product officer at Cisco Systems — one of the few dozen organizations with early access — described the model as "potent" and said Cisco had used it to proactively identify vulnerabilities in its own systems. „I actually feel like there will be a world that is pre-Mythos and post-Mythos.” — Jeetu Patel via Bloomberg Business Anthropic co-founder Jack Clark confirmed on Monday that the company had been discussing Mythos with the Trump administration, even after the Pentagon severed its business relationship with the firm.

Anthropic was founded in 2021 by former members of OpenAI, including siblings Daniela Amodei and Dario Amodei. The company has historically been among the most vocal in the AI industry about warning of potential dangers from advanced models. The dispute between Anthropic and the Pentagon arose over the company's refusal to allow its models to be used for mass surveillance or to develop fully autonomous weapons — a position the Defense Department argued was unduly restrictive. The designation of Anthropic as a supply chain risk was described as historically unusual, as such designations had previously been reserved for foreign companies.

Global finance leaders alarmed as Treasury summons Wall Street The model's potential impact on financial stability has drawn concern from senior policymakers well beyond Washington. Treasury Secretary Scott Bessent summoned the leaders of some of the largest U.S. banks to discuss the cyber risks posed by Mythos, a meeting that Bloomberg News reported helped trigger a state of high alert among officials attending the spring meetings of the International Monetary Fund and the World Bank in Washington. European Central Bank President Christine Lagarde addressed the model's risks directly in a television interview. „If it falls in the wrong hands, it could be really bad.” — Christine Lagarde via Bloomberg Business Bank of England Governor Andrew Bailey, speaking at an event at Columbia University in New York, described waking up to find that Anthropic "may have found a way to crack the whole cyber risk world open." The broader competitive landscape in artificial intelligence also shifted this week, with OpenAI announcing a pivot toward business users and a new model aimed at "high-value professional work," as the company faces intensifying competition from Anthropic for corporate clients. 852 (billion USD) — OpenAI's reported valuation OpenAI reported that business customers accounted for approximately 40 percent of its revenue, up from around 20 percent when its chief financial officer Sarah Friar joined the company in 2024, according to the Associated Press.