IT Brief US - Technology news for CIOs & IT decision-makers
United States
Rubrik uses Anthropic Mythos to probe software flaws

Rubrik uses Anthropic Mythos to probe software flaws

Tue, 2nd Jun 2026 (Yesterday)

Rubrik has gained access to Anthropic's Mythos Research Preview through Project Glasswing, a programme that gives selected operators of critical software infrastructure access to advanced AI models for vulnerability discovery and remediation.

The cybersecurity company said it will use the technology to identify, assess and address potential software vulnerabilities across its enterprise platform and product portfolio. The move forms part of a broader effort to strengthen cyber resilience as AI accelerates the pace at which software flaws can be discovered and potentially exploited.

Anthropic is providing access to Mythos Research Preview to organisations responsible for building or maintaining critical software infrastructure. The objective is to help participants identify and fix vulnerabilities before they can be used in attacks.

Vulnerability testing

Rubrik said it will apply the model directly to its codebase to help uncover potential weaknesses and support remediation efforts before vulnerabilities can be exploited.

"It's important for us to be a part of Anthropic's Glasswing project as we work to speed trusted AI transformation for our customers, which includes both trust in our software and in our ability to deliver preemptive recovery and resilience," said Bipul Sinha, CEO, Chairman and Co-founder, Rubrik.

"By putting Anthropic's model to work directly on our code, we are proactively purging potential vulnerabilities before they can ever be leveraged against us," said Sinha.

The company said the initiative aligns with its focus on cyber resilience, data protection and recovery. Rubrik's platform is designed to secure, monitor and recover data, identities and workloads across cloud environments.

Faster discovery

The collaboration reflects growing concern across the cybersecurity industry about the effect of advanced AI models on vulnerability research.

According to Rubrik, recent developments have demonstrated that AI systems are capable of identifying software vulnerabilities at speeds that significantly reduce the time between discovery and exploitation.

In commentary accompanying the announcement, Sinha said the cybersecurity sector is entering a period where vulnerability discovery is accelerating rapidly through AI-driven analysis.

He pointed to findings disclosed by Anthropic earlier this year that showed an advanced frontier model identifying thousands of software vulnerabilities across browsers and operating systems. Some of those vulnerabilities had reportedly remained undetected for extended periods.

Rubrik argues that this shift is changing assumptions about incident response and recovery planning. The company believes organisations must focus on preparedness before attacks occur rather than relying solely on post-incident remediation.

Recovery focus

Rubrik said access to Mythos Research Preview supports its strategy of pre-emptive recovery and resilience.

The company maintains that the shrinking gap between vulnerability discovery and exploitation requires organisations to strengthen recovery capabilities alongside traditional prevention measures.

"Anthropic is helping its partners plan ahead, to be preemptive, to vet their systems and apply fixes that position them best for future recovery and resilience. We could not agree more with this approach," said Sinha.

Rubrik said the model will help it identify vulnerabilities within its environment, strengthen platform security and reduce risk for customers.

Industry sharing

As a participant in Project Glasswing, Rubrik also plans to contribute findings from its testing activities back to the broader security community.

The company said sharing insights from vulnerability discovery and remediation efforts could help improve preparedness across the industry as AI-assisted security research continues to develop.

Rubrik views the initiative as part of a wider transition in cybersecurity, where AI technologies are increasingly used for both defensive and offensive purposes.

"Anthropic has shown us the start of the next evolution in cybersecurity, one that is quickly moving. Rubrik is here, we are grateful to be part of Project Glasswing, and we are ready for resilience," said Sinha.