NYSE deployed Anthropic’s Claude Mythos Preview to find and patch cybersecurity flaws

  • NYSE President Lynn Martin said the exchange used Claude Mythos Preview to fix cybersecurity weaknesses.
  • Anthropic’s partners reported bug discovery rates more than tenfold higher than previous methods.
  • Project Glasswing expanded to roughly 150 organizations across healthcare, utilities and communications by early June.

The New York Stock Exchange deployed Anthropic’s Claude Mythos Preview to identify and remediate cybersecurity vulnerabilities across its systems, NYSE President Lynn Martin told the House Financial Services Committee on September 2, 2026. Some flaws had apparently remained in the codebase for decades, although Martin did not disclose their severity, remediation timeline or cost. Anthropic launched Project Glasswing on April 7, 2026, granting selected organizations access to the model for deep code analysis and vulnerability detection. Amazon Web Services, Apple, Google and Microsoft were among its initial partners, while Intercontinental Exchange formally joined on June 3 and deployed the model across all its platforms. Participants reported bug discovery rates more than ten times higher than with previous methods, and the broader program surfaced thousands of critical vulnerabilities. Anthropic charges $25 per million input tokens and $125 per million output tokens, and has committed up to $100 million in usage credits plus $4 million in security donations. By early June, the initiative had expanded to roughly 150 additional organizations in healthcare, power utilities and communications. Intercontinental Exchange shares fell 1.39% on the day of Martin’s testimony.

The information on this website is generated using AI and we cannot guarantee its accuracy. Please use it as reference information only.