NEW YORK CITY, New York: Microsoft is expanding its cybersecurity toolkit by integrating advanced artificial intelligence models, including Anthropic's Claude Mythos Preview, into its software development processes to detect better and fix vulnerabilities.

The company said the models will be embedded within its Security Development Lifecycle (SDL), allowing engineers to identify potential security flaws earlier and respond more quickly.

In a blog post, Microsoft said incorporating these tools would help "identify vulnerabilities and develop fixes faster" during the initial stages of software development.

Mythos, introduced by Anthropic on April 7, has already uncovered "thousands" of major vulnerabilities across operating systems, web browsers, and other software, according to experts. Its ability to write and analyse code at a high level gives it a powerful edge in detecting weaknesses and potentially exploiting them.

Anthropic said the current version, Claude Mythos Preview, will initially be rolled out to a limited group of companies as part of its "Project Glasswing" initiative. The controlled program allows major technology firms, including Microsoft, Amazon, and Apple, to use the tool to search for cybersecurity vulnerabilities.

Microsoft said it evaluated Mythos using its own open-source benchmark designed for real-world detection engineering tasks, adding that the "results showed substantial improvements relative to prior models."

The push comes as governments and industries globally race to understand and adopt the new technology. U.S. President Donald Trump's administration, central banks, and private sector firms are all seeking to adapt to Mythos and its potential to both strengthen cybersecurity and accelerate complex cyberattacks.

Experts say tools like Mythos could reshape how vulnerabilities are discovered and addressed, while also raising concerns about how such powerful capabilities might be used if widely deployed.