Anthropic is set to collaborate with partners next week to expand Project Glasswing, which is focused on enhancing security through vulnerability discovery. This initiative is bolstered by the involvement of major organizations that have participated in the project, with reports indicating that most partners have identified hundreds of vulnerabilities utilizing the AI tools provided for their defensive security efforts.

Anthropic: Anthropic is an AI research company focused on developing safe and capable frontier models, including the Claude family. It recently launched Project Glasswing, a cybersecurity initiative that provides select partners with access to its advanced Claude Mythos Preview model for defensive vulnerability detection and remediation in critical software. The company is actively expanding this collaborative effort with industry partners to address emerging AI-driven security challenges.
Glasswing: Project Glasswing is Anthropic’s industry-wide initiative aimed at securing critical software infrastructure against advanced cyber threats using AI capabilities. Launched in April 2026, it unites technology and security leaders to scan and fix vulnerabilities, with the project now moving into an expansion phase involving additional partners and shared findings. Participating organizations have reported significant success in identifying long-hidden flaws through the associated AI tools.

Initiative Expansion: Anthropic plans to work with partners to expand Project Glasswing in the coming week, building on initial partner contributions to vulnerability discovery.
Industry Collaboration: Recent updates show ongoing involvement from major organizations strengthening their security efforts through continued participation in Project Glasswing.
Vulnerability Findings: Most partners in the Glasswing coalition have each identified hundreds of vulnerabilities using the provided AI tools for defensive security work.