Goldman Sachs and Citigroup Are Testing Anthropic's Mythos AI Model Built for Finance

Neural network AI visualization overlaid on Wall Street finance data and bank vault - Anthropic Mythos banking AI

Major Wall Street banks including Goldman Sachs and Citigroup are internally testing a new AI model from Anthropic purpose-built for the financial sector, according to sources familiar with the matter. The model, called Mythos, is part of Anthropic's push into industry-specific AI applications under a project codenamed Glasswing — with JPMorgan Chase named as a key participant.

What Is Mythos?

Mythos is Anthropic's specialized AI model trained on financial data and designed for use cases specific to banking and capital markets. Unlike Claude, Anthropic's general-purpose AI assistant, Mythos is tailored for tasks such as financial analysis, risk assessment, regulatory compliance, and investment research.

The model represents a significant strategic shift for Anthropic, which has largely focused on general AI capabilities. By developing a finance-specific model, the company is directly competing with financial AI offerings from OpenAI and a growing number of specialized fintech AI startups.

Project Glasswing and Bank Involvement

Sources say Project Glasswing — the internal codename for the banking AI initiative — has attracted interest from multiple top-tier financial institutions. Goldman Sachs and Citigroup have been conducting internal testing, while JPMorgan Chase is the only bank officially named in connection with the program.

The banks are evaluating Mythos for a range of use cases, from automating analyst workflows to generating financial reports and flagging regulatory anomalies. Financial institutions have been eager to adopt AI for productivity gains but cautious about security, accuracy, and regulatory risk.

Why This Matters for the AI Industry

The launch of Mythos signals a broader trend: the next phase of AI competition is moving from general-purpose models to domain-specific ones. Healthcare, legal, and finance are emerging as the most lucrative verticals, and major AI labs are racing to secure long-term enterprise contracts in these sectors.

Anthropic's recent surge in business adoption — reaching 30.6% of US businesses on the Ramp platform in March 2026 — has given the company the credibility and revenue to invest in vertical expansions like Mythos.

The Bottom Line

Anthropic is betting that the future of enterprise AI lies in specialized models that understand the nuances of specific industries. If Mythos gains traction on Wall Street, it could become one of the company's most valuable products — and a template for similar launches across healthcare, law, and government sectors.