Anthropic is enabling an additional 150 organizations around the world to access Mythos, an artificial intelligence model aimed at finding cybersecurity vulnerabilities that the company considered too dangerous to be made available to the general public.
Anthropic is expected to announce this Tuesday (2) the increase in the number of users, bringing the total number of groups with access to the system to around 200. According to the company, the new organizations are in 15 countries and operate in sectors such as energy, health and communications.
While Anthropic declined to name the new participants, it said the group includes companies and nonprofits that produce widely used essential programming code. People familiar with the matter said the European Union’s cybersecurity body will be given access to the software.
Study abroad
Upgrade your career!
The movement occurs after Anthropic confidentially filed documentation for an initial public offering (IPO), overtaking rival OpenAI in this race. In a post published on the company’s blog on Monday, the company said it has not yet defined how many shares it intends to offer nor the price.
Previously seen as an underdog compared to OpenAI, Anthropic raised US$65 billion in a funding round last week, at a valuation of US$965 billion, including the new investment. It was the first time its rating surpassed OpenAI’s. Anthropic’s advances in programming and cybersecurity capabilities have been disrupting markets and attracting new corporate clients.
Mythos has become one of the San Francisco-based company’s focuses in recent months. Anthropic claims that the software is capable of identifying and exploiting vulnerabilities “in all major operating systems and in all major browsers when instructed by the user to do so.”
Continues after advertising
Therefore, the company initially limited access to Mythos to a small group of large technology companies and Wall Street institutions, within an initiative called Project Glasswing, launched in April.
Since the launch of Mythos, the software has been used to find more than 10,000 serious vulnerabilities, according to Anthropic. The company said in recent days that it plans to widely release new AI models in the coming weeks with cybersecurity capabilities comparable to Mythos. Anthropic said it has made “rapid progress” in developing the security safeguards necessary to achieve this.
© 2026 Bloomberg L.P.
