Arm Launches Metis AI Framework for Enhanced Software Security

• •Arm's Metis framework enhances vulnerability detection in complex software systems.
• •The framework is already used in over 130 internal projects and will expand by late 2026.
• •Metis reduces false positives by about 50%, improving developer efficiency.

Arm has introduced Metis, an open-source AI-driven security framework aimed at improving vulnerability detection in complex software systems. Developed by Arm’s product security team, Metis is currently operational across over 130 internal projects, with plans for broader adoption by late 2026. The framework addresses limitations of traditional static analysis tools by utilizing advanced AI techniques to identify vulnerabilities that span multiple components and layers. Internal benchmarks indicate Metis can achieve up to ten times higher true positive detection rates and reduce false positives by approximately 50%. Built on a retrieval-augmented generation architecture, Metis combines large language models with project-specific knowledge for contextual analysis. It can validate findings from its own analysis and external security tools, distinguishing real vulnerabilities from false alarms. The framework is designed to support developers by identifying issues earlier and improving overall product security and performance.