As part of Anthropic’s Project Glasswing expansion, BeyondTrust gains access to Claude Mythos Preview.
In modern enterprise security, identity has fast become the primary operational perimeter, and administrative privilege is the ultimate target inside it. As AI drastically accelerates both software development and software exploitation, the window between discovering a software vulnerability and it being actively weaponized has collapsed from months to mere seconds.
Recognizing this rapid shift in threat velocity, BeyondTrust has announced its selection into an expanded cohort of Anthropic’s Project Glasswing. BeyondTrust is now among a group of vendors whose codebases maintain critical digital infrastructure across global commerce, government, healthcare, and essential services. As organizations increasingly rely on privileged identities to operate and secure those environments, protecting the software that governs privilege has become an essential part of cyber resilience.
“Millions of the world’s most important workloads are protected by privilege controls that we build and maintain. That is a responsibility we take seriously,” said Janine Seebeck, Chief Executive Officer, BeyondTrust. “As AI reshapes both software development and cyber defense, the organizations responsible for securing critical infrastructure must continuously raise the bar. Project Glasswing gives us an opportunity to do exactly that by continuing to strengthen the security and resilience of the software our customers rely on to protect all human, machine, and agentic identities.”
The rapid growth of human, machine, and agentic identities is creating unprecedented complexity across enterprise environments. AI agents are now one of the fastest-growing yet least-governed classes of privileged identities, often operating with access to sensitive systems, data, and business processes. They join human administrators, service accounts, and workload identities as increasingly attractive targets for attackers. For the operators of critical infrastructure, resilience will increasingly depend on securing the identities and privileges that connect every user, system, application, and AI-driven process.
BeyondTrust Joins Anthropic’s Project Glasswing
As part of Anthropic’s Project Glasswing expansion, BeyondTrust gains access to Claude Mythos Preview, a frontier AI model that has already demonstrated significant defensive potential, helping program partners surface more than 10,000 high and critical severity vulnerabilities.
“The threats ahead are bigger than any one vendor, and the response has to be shared,” said Marc Maiffret, Chief Technology Officer, BeyondTrust. “We are honored to stand with Anthropic and the other members of Project Glasswing, applying Mythos to our own code to further strengthen the security of the products our customers depend on, and doing our part in a defense no one can mount alone.”
Gated Defense: Inside Anthropic’s Selection Criteria
BeyondTrust’s inclusion comes as part of a highly controlled, strategic expansion of Project Glasswing announced by Anthropic on June 2, 2026. Moving past its initial cohort of roughly 50 localized hyperscalers and government agencies, Anthropic has extended Mythos access to an elite group of 150 organizations across 15 countries.
The selection process is strictly gated. Because Anthropic has withheld Mythos from general public release due to safety risks, it has actively hand-picked infrastructure operators and cornerstone software vendors. For most chosen partners in this cohort, Anthropic estimates a major breach could impact upwards of 100 million people. This makes BeyondTrust, which commands the privilege pathways for 75 of the Fortune 100, an important addition to the program.
BeyondTrust will use Claude Mythos Preview defensively to identify, validate, and remediate potential software vulnerabilities across its product portfolio and the BeyondTrust Pathfinder Platform.
To learn more about BeyondTrust’s inclusion in Project Glasswing, read the full blog post here.
