AWS and Anthropic Partner to Boost Cybersecurity with Claude Mythos AI

Key Takeaways

• AWS and Anthropic have launched Claude Mythos Preview, an AI model designed to identify and patch software vulnerabilities at scale.
• The new AI model is part of Anthropic’s Project Glasswing and is currently in a gated research preview through Amazon Bedrock.
• Access to Claude Mythos Preview is restricted to an allow-list of critical organizations and open-source maintainers.
• AWS has also made its Security Agent generally available, offering continuous, AI-driven penetration testing across multi-cloud and on-premises environments.

In a significant move to counter the rapidly escalating landscape of cyber threats, Amazon Web Services (AWS) and Anthropic have announced a strategic collaboration aimed at leveraging next-generation artificial intelligence for enhanced cybersecurity. This partnership introduces advanced AI capabilities designed to secure critical software infrastructure globally.

As part of Anthropic’s newly unveiled Project Glasswing, a specialized AI model named Claude Mythos Preview is now entering a restricted release phase. This initiative seeks to bolster the security posture of essential software systems worldwide.

AWS has long integrated machine learning into its extensive infrastructure protection strategies. The company’s internal AI-powered log analysis capabilities have dramatically reduced security review times, cutting an average of six hours down to merely seven minutes. Furthermore, AWS successfully thwarted over 300 million malicious S3 encryption attempts in 2025 alone by analyzing more than 400 trillion network flows daily. These sophisticated defense mechanisms are now being extended to enterprise clients.

Claude Mythos Enters Preview

Claude Mythos Preview represents a substantial advancement in AI reasoning, specifically engineered for complex software coding and cybersecurity applications. Its primary function is to efficiently discover and remediate vulnerabilities on an extensive scale, requiring minimal human intervention from engineers.

Recognizing the inherent risks associated with AI models capable of generating functional exploits, AWS and Anthropic are implementing a highly cautious distribution strategy. Claude Mythos Preview is exclusively accessible through a gated research preview via Amazon Bedrock.

Access is strictly limited to an approved list of organizations, prioritizing companies critical to internet infrastructure and major open-source maintainers whose software influences hundreds of millions of users. Participating organizations testing the model on Amazon Bedrock benefit from robust, enterprise-grade security protocols, allowing teams to explore the AI’s capabilities without exposing production assets to undue risk. Key protections include:

• Customer-managed data encryption to ensure privacy.
• Strict Virtual Private Cloud (VPC) isolation for secure workloads.
• Automated Reasoning safeguards providing 99% accuracy against factual hallucinations.

Autonomous Penetration Testing

Concurrent with the Claude Mythos announcement, AWS has officially made its new AWS Security Agent generally available. This tool is set to revolutionize enterprise penetration testing, transforming it from a periodic, manual process into a continuous, on-demand capability.

Operating around the clock, the Security Agent employs specialized AI to autonomously identify and validate vulnerabilities across AWS, Azure, Google Cloud Platform (GCP), and on-premises environments. Unlike conventional security scanners that often produce unverified alerts, the AWS Security Agent actively attempts to exploit vulnerabilities using targeted attack chains.

Upon successfully confirming a legitimate risk, it furnishes security teams with comprehensive, actionable documentation. This includes:

• Standardized CVSS risk scores to define the threat level.
• Application-specific severity ratings to contextualize the danger.
• Detailed reproduction steps illustrating how the exploit functions.
• Practical remediation suggestions for immediate fixes.

As nation-state actors and ransomware groups increasingly leverage AI to scale their malicious operations, defensive strategies must adapt proactively. By integrating Anthropic’s advanced threat detection capabilities with autonomous cloud security agents, AWS aims to empower organizations to establish resilient defenses before emerging threats can materialize.

Disclaimer: HackersRadar reports on cybersecurity threats and incidents for informational and awareness purposes only. We do not engage in hacking activities, data exfiltration, or the hosting or distribution of stolen or leaked information. All content is based on publicly available sources.