Anthropic's Claude Mythos Powers Critical Infrastructure in 15+ Countries

Anthropic announced on Tuesday the significant expansion of Project Glasswing, its collaborative industry program utilizing artificial intelligence to identify and remediate critical software vulnerabilities. This initiative will now extend its reach to approximately 150 new organizations spanning over 15 countries.

This news follows Anthropic's confidential filing for an initial public offering just one day prior. The IPO pursuit comes on the heels of a substantial $65 billion funding round, which valued the company at nearly $1 trillion.

At the core of Project Glasswing is Anthropic’s Claude Mythos, a model the AI firm describes as its most powerful to date. Claude Mythos is engineered to detect thousands of zero-day vulnerabilities within weeks. In early April, Anthropic granted 50 initial partners, including the U.S. government, access to Claude Mythos Preview to scan their codebases for potential vulnerabilities and security flaws.

The newly expanded roster of organizations with access to Mythos now encompasses vital sectors such as power, water, healthcare, communications, and hardware. These industries, according to Anthropic, were not adequately represented in the initial cohort. Many of the newly included entities are companies or non-profits managing codebases that are critical to the operations of other organizations and governments, as highlighted in a company blog post.

"What each partner has in common is that a successful attack on their codebase could be catastrophic," Anthropic stated. The company further emphasized the potential impact, estimating that "For most partners, we estimate that a major attack could affect more than 100 million people, with important ramifications for both global and national security."

The expanded group includes organizations situated in countries considered friendly to the U.S., according to The Financial Times, which cited an informed source. These nations include Australia, Canada, France, Germany, Italy, Switzerland, the Netherlands, Spain, Belgium, Sweden, India, Japan, New Zealand, and South Korea.

The Financial Times also identified several specific organizations granted access to Mythos. These include Okta, a U.S.-based identity and security management provider; South Korean corporations Samsung, SK Hynix, and SK Telecom; NATO, the U.S.-led military alliance headquartered in Brussels; and ENISA, the European Union’s cybersecurity agency.

TechCrunch has reportedly sought confirmation from Anthropic regarding these details.

Anthropic has indicated its expectation that other AI companies will soon develop models with capabilities comparable to Mythos Preview. This foresight is driving the firm’s accelerated efforts to establish robust safeguards within Project Glasswing.

In a related development, since the introduction of Mythos, competitor OpenAI has launched its own cybersecurity-focused model, GPT-5.5-Cyber, which is currently being rolled out to a broad group of partners for testing.