According to Anthropic, the latest cohort includes organisations whose services are essential to millions of people and whose disruption could have far-reaching consequences. The company said the newly added partners share a common characteristic: a successful cyberattack on their software codebases could have catastrophic effects. For most participating organisations, Anthropic estimates that a major attack could impact more than 100 million people, carrying serious implications for both national and global security.
Initially, Project Glasswing included only around 50 partners, primarily based in the United States, allowing Anthropic to evaluate the model’s capabilities before the technology could potentially be exploited by malicious actors. The latest expansion marks the first major extension of the programme beyond the United States and the United Kingdom.
Announcing the expansion in its latest statement, Anthropic said it is extending Project Glasswing following several weeks of close collaboration with programme partners, the cybersecurity industry, open-source software maintainers, and the United States government. The company stated that each newly selected organisation will be required to meet specific security requirements before receiving access to the model.
Anthropic said the organisations joining the programme are located in more than 15 countries and that most provide critical infrastructure supporting populations far beyond their national borders. The company described the expansion as another step toward its long-term objective of using artificial intelligence to improve software security while helping the cybersecurity industry adapt to the profound changes advanced AI systems could bring to the field.
Claude Mythos Preview is Anthropic’s specialised artificial intelligence model developed for advanced cybersecurity research, software code analysis, and vulnerability discovery. The company has not made the model publicly available, citing concerns that its capabilities could be misused. According to Anthropic, Mythos excels at analysing massive software codebases and identifying security flaws that may have remained undetected for years.
The company claims the model has demonstrated an ability to uncover thousands of software vulnerabilities and generate proof-of-concept exploit code at a speed significantly faster than conventional methods. These capabilities have drawn attention from cybersecurity professionals and regulators worldwide. Anthropic has repeatedly stated that public release of the model is being withheld because malicious actors could potentially use it to identify and exploit critical software weaknesses.
Project Glasswing was launched alongside Claude Mythos in April 2026 as a collaborative cybersecurity initiative designed to strengthen digital security without broadly distributing the model. Instead of public access, Anthropic initially provided the technology to a carefully selected group of approximately 50 partners, including major technology companies, cybersecurity firms, financial institutions, and open-source organisations.
Early participants included Apple, Microsoft, Google, Amazon Web Services, Nvidia, Cisco, Broadcom, CrowdStrike, JPMorgan Chase, and the Linux Foundation. The programme’s primary objective is to deploy Mythos to identify, analyse, and help remediate vulnerabilities across critical digital infrastructure before they can be exploited by cybercriminals or state-sponsored threat actors.
Anthropic stated that organisations participating in Project Glasswing have already identified more than 10,000 high-severity or critical vulnerabilities through the deployment of the model. The initiative also brings together governments, cybersecurity experts, infrastructure providers, and open-source software maintainers to reinforce digital defenses across sectors including banking, communications, and public services.
India’s inclusion in the expanded programme places the country among a limited group of trusted participants granted access to one of the world’s most advanced cybersecurity artificial intelligence systems. The development underscores the growing role of artificial intelligence in protecting critical infrastructure and highlights increasing international cooperation to address evolving cyber threats before they can cause large-scale disruption.
About The Author
Comment List