Anthropic's new Mythos Preview model is moving from a tightly controlled security program into early testing at major banks, according to the company's own launch materials and a subsequent Bloomberg report.

What Anthropic confirmed

On April 7, Anthropic launched Project Glasswing, a defensive security initiative built around Mythos Preview, an unreleased frontier model the company says is unusually strong at finding and exploiting software vulnerabilities. Anthropic named JPMorganChase among Glasswing partners, alongside companies including AWS, Apple, Broadcom, Cisco, CrowdStrike, Google, Microsoft, NVIDIA, Palo Alto Networks, and the Linux Foundation.

Anthropic also said it extended Mythos access to more than 40 additional organizations that build or maintain critical software infrastructure, while committing up to $100 million in usage credits and $4 million in direct donations to open-source security groups.

What appears to be new

Bloomberg reported on April 10 that other large Wall Street firms, including Goldman Sachs, Citigroup, Bank of America, and Morgan Stanley, are also testing Mythos internally or expecting access shortly. Bloomberg further reported that Trump administration officials encouraged banks to use the model to identify vulnerabilities.

The cautious takeaway is that Mythos is no longer just an Anthropic research story. If Bloomberg's reporting is accurate, the model is already being treated as infrastructure-grade security tooling inside large financial institutions, even while Anthropic keeps general availability tightly limited.