US officials reportedly push banks to test Anthropic’s Mythos

This week, Scott Bessent, the US Treasury Secretary, and Jerome Powell, the chair of the Federal Reserve—the US central bank—summoned bank executives to a meeting. According to Bloomberg, the officials encouraged the executives to use Anthropic’s new Mythos model to detect vulnerabilities in their systems.

The government’s encouragement comes as a broader group of financial institutions begins evaluating the technology. While JPMorgan Chase was the only bank listed as an initial partner organization with access to the model, several other major banks are reportedly testing Mythos as well. These institutions include:

• Goldman Sachs
• Citigroup
• Bank of America
• Morgan Stanley

Anthropic announced the Mythos model this week but stated that it would be limiting access for now. According to the company, this restriction is in part because the model is highly effective at finding security vulnerabilities, despite not being specifically trained for cybersecurity.

This push for adoption by US officials occurs alongside an active legal dispute. Anthropic is currently in court battling the Trump administration over a Department of Defense designation that labels the AI company a “supply-chain risk.” This designation was issued after negotiations broke down regarding Anthropic’s efforts to limit how the US government can use its AI models.

At the same time, the model is drawing scrutiny outside the United States. The Financial Times reports that U.K. financial regulators are also discussing the potential risks posed by Mythos.

Why it matters

The situation highlights a complex dynamic where US officials are encouraging banks to adopt Anthropic’s Mythos model while the company remains in a legal battle with the Trump administration over a Department of Defense “supply-chain risk” designation.