Anthropic is expanding its Project Glasswing initiative, which focuses on using artificial intelligence to identify and fix critical software vulnerabilities. According to a statement released on Tuesday, nearly 150 new organizations from over 15 countries have joined the project. This news follows reports that Anthropic has filed confidentially for an IPO with a $1 trillion valuation, following a $65 billion investment round. This is reported by Techcrunch.com reports.
At the heart of Project Glasswing is the Claude Mythos model. The company calls this model its most powerful development, capable of identifying thousands of "zero-day" vulnerabilities within weeks. In early April, Anthropic gave 50 initial partners, including the US government, access to the Claude Mythos Preview version to scan their codebases for security flaws.
Today, the list of organizations with access to the Mythos model covers sectors such as energy, water supply, healthcare, telecommunications, and hardware. Anthropic noted in a blog post that these partners share a common trait: a successful attack on their codebases could have catastrophic consequences for global and national security.
The Financial Times reports that the expanded group includes organizations from US-allied countries such as Australia, Canada, France, Germany, Italy, India, Japan, and South Korea. Specifically, the US company Okta, South Korean giants Samsung, SK Hynix, and SK Telecom, as well as NATO and the EU cybersecurity agency ENISA, have gained access to the model.
Anthropic expects other AI companies to develop models at the Mythos Preview level soon, which is why it is rushing to establish security measures within Project Glasswing. Competitor OpenAI has already introduced its cybersecurity-focused GPT-5.5-Cyber model and distributed it to a large group of partners for testing.
Read “Zamin” on Telegram!
