US regulators push for adoption of Anthropic's new tool, prompting major Wall Street banks to enter "self-inspection mode"

Glonghui, April 12 — The latest news shows that, driven by U.S. government officials, major Wall Street banks are preparing to internally test Anthropic's newly launched Mythos model for detecting system vulnerabilities. Earlier this week, Anthropic announced on its official website the launch of the "Glasswing" project: leveraging the company’s newly trained advanced model, "Claude Mythos Preview", to provide technical support for the security of critical software infrastructure. According to the statement, Mythos Preview has already identified thousands of high-risk vulnerabilities, covering almost all mainstream operating systems and web browsers. The reason for launching the "Glasswing" project is the company's belief that it has the potential to reshape the cybersecurity landscape. The statement mentioned that Amazon, Apple, Broadcom, Cisco, Google, Microsoft, Nvidia and other companies have already begun testing, and Anthropic has provided $100 million in usage credits for this. Among banks, only JPMorgan is participating. The latest updates indicate that, in addition to JPMorgan, other major financial institutions have also obtained permission to use the model within the next few days. According to sources familiar with the matter, Goldman Sachs, Citi, Bank of America, Morgan Stanley and others are already conducting internal technical tests.