Anthropic Expands Project Glasswing to Approximately 150 New Organizations

Anthropic announced on June 2, 2026, that it is significantly expanding Project Glasswing, its initiative to secure critical software using the Claude Mythos Preview model. After launching in early April with roughly 50 initial partners who have since identified more than 10,000 high- or critical-severity security flaws, Anthropic is now extending the program to approximately 150 new organizations across more than 15 countries. The new cohort includes vendors and operators in power, water, healthcare, communications, and hardware sectors—many of whom maintain codebases relied upon globally—where a successful attack could affect more than 100 million people.

The expansion reflects Anthropic’s long-term goal of using AI to make all software more secure and to prepare the industry for a near-term future in which multiple AI companies are expected to release Mythos-class cyber-capable models within 6 to 12 months, potentially without safeguards against misuse. Anthropic outlined a twofold role for itself: providing wide access to defensive models, tools, and common infrastructure; and shifting support from vulnerability discovery toward disclosure, patching, and deployment of fixes.

To support defenders beyond the Glasswing partnership, Anthropic released Claude Security, a product that uses its latest public frontier models such as Claude Opus 4.8 to scan codebases and suggest patches, and is offering its internal Glasswing vulnerability-finding tools on request to trusted security teams. Partners are already using Mythos Preview to write patches, conduct pre-release checks, perform penetration testing, automate threat detection and response, and rebuild legacy codebases in memory-safe languages. Anthropic noted that the current bottleneck in cybersecurity is verifying and patching vulnerabilities at scale, and said it is in discussions to scale open-source vulnerability review while developing best practices for disclosure to maintainers.

General access to Mythos-level capabilities remains a future goal pending the development of robust safeguards that prevent misuse of offensive cyber capabilities. Anthropic plans further expansions of Project Glasswing and its Cyber Verification Program to reach additional essential infrastructure providers, critical open-source maintainers, and safety testers in the United States and overseas.

Source: Anthropic News