Per a Bloomberg report, U.S. Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell convened a closed-door meeting with leading bank executives this week. During the gathering, the top U.S. economic officials encouraged industry leaders to leverage Anthropic’s newly released Mythos AI model to detect security and systemic vulnerabilities across their operations.

While JPMorgan Chase remains the only major bank officially listed as an initial partner granted early access to the model, multiple other large Wall Street institutions — including Goldman Sachs, Citigroup, Bank of America, and Morgan Stanley — are already conducting internal trials of Mythos, per the outlet’s sources.

Anthropic formally announced the launch of the Mythos model this week, but confirmed that access will remain restricted for the immediate future. The company shared one core reasoning for the limited rollout: even though the model was not specifically trained for cybersecurity use cases, it is far more effective at uncovering unpatched security gaps than many expected. That said, some industry observers have pushed back on this framing, arguing the limited access policy is either overblown hype or a calculated, smart enterprise sales strategy to drive demand.

The push from top U.S. officials to embrace Anthropic’s technology is especially surprising, given the AI company’s ongoing high-stakes legal battle with the Trump administration. The conflict stems from the U.S. Department of Defense’s decision to classify Anthropic as a national supply chain risk, a designation issued after negotiations between the firm and the agency collapsed. Talks fell apart over Anthropic’s push to place strict limits on how its AI models can be used by the U.S. government.

Separately, the Financial Times reports that U.K. financial regulators are also holding internal discussions about the potential risks that the Mythos model poses to the country’s financial sector.