Anthropic, the creator of the Claude AI models, has expanded the availability of its powerful cybersecurity-focused Mythos AI model to over 15 additional countries. This move comes months after the company decided not to release the model to the general public due to concerns about potential misuse. The newly added countries include Australia, Canada, France, Germany, Italy, Switzerland, the Netherlands, Spain, Belgium, Sweden, India, Japan, New Zealand, and South Korea.
Expansion of Project Glasswing
In an official announcement, Anthropic stated that it is significantly expanding Project Glasswing, an initiative that provides selected organizations with access to Claude Mythos Preview. This advanced AI model is designed to identify software vulnerabilities and strengthen cyber defenses. Initially rolled out to a limited number of companies in April, the launch had raised concerns among banks and financial institutions regarding cybersecurity threats.
According to Anthropic, around 50 organizations were initially given access to the model in April. Since then, these partners have used Mythos to scan their software and collectively discovered more than 10,000 high- and critical-severity security flaws. The program has now been extended to approximately 150 additional organizations, provided they meet the company's security requirements.
“We’re now expanding Project Glasswing. Following several weeks of close collaboration with our Project Glasswing partners, the security industry, open-source software maintainers, and the US government, we’re extending the partnership to approximately 150 new organizations. Each one will need to meet our security requirements before they gain access,” the company said.
“The organizations in this new group are based in more than 15 countries, and most provide critical infrastructure to many more. (In the future, we intend to expand our geographical reach much further.) The group covers several industries that weren’t well represented in our initial cohort, such as power, water, healthcare, communications, and hardware. And many of the new partners are vendors—companies or nonprofits that maintain codebases that are relied upon by lots of other organizations around the world, including governments,” it added.
Critical Infrastructure Focus
Anthropic emphasized that the new group includes organizations from more than 15 countries and covers industries such as power, water, healthcare, communications, hardware, and critical infrastructure. Many of the newly added participants are software vendors and nonprofit organizations that maintain code used by governments, businesses, and institutions worldwide. “What each partner has in common is that a successful attack on their codebase could be catastrophic,” Anthropic said. The company added that for most participants, a major cyberattack could potentially affect more than 100 million people.
Why Mythos Remains Restricted
Anthropic reiterated that it still does not plan to release Mythos-level capabilities for general use because strong safeguards against misuse have not yet been developed. The company warned that advanced cyber-capable AI systems could soon become widely available across the industry. “Within 6 to 12 months, we expect that many other AI companies will have Mythos-class models, and they could release them without safeguards that prevent misuse,” Anthropic said. The AI startup believes such models could make cyberattacks more frequent and unpredictable if used by malicious actors. Because of these risks, Anthropic is limiting access to trusted organizations while working on stronger safety protections.
