The new model, MAI-Cyber-1-Flash, powers Microsoft’s existing MDASH software vulnerability harness. According to Microsoft AI CEO Mustafa Suleyman, the model integrates with GPT 5.4 to outperform rival systems like Gemini and Mythos 5 on the Cyber Gym benchmark. The company plans to move these capabilities into production immediately, with a public preview release scheduled for November 3.
In section Startups & Technology
Microsoft Enters AI Security Market with MAI-Cyber-1-Flash and Perception
At a San Francisco event, Microsoft unveiled its first cybersecurity-specialized model, MAI-Cyber-1-Flash, alongside Perception, an agent-based platform designed to automate vulnerability detection and remediation. The move marks a direct challenge to established AI security offerings from competitors like Anthropic, OpenAI, and Google in an increasingly crowded enterprise landscape.
Perception functions by deploying specialized agentic teams—red, blue, and green—to handle the lifecycle of a security threat. Red teams simulate potential attacks to identify vulnerabilities, blue teams triage existing bugs, and green teams execute corrective actions. Dave Weston, the platform's lead engineer, stated that the system compresses tasks that previously required hours of manual cross-departmental labor into minutes of automated response. By offering integrated detection, posture management, and direct code fixes, Microsoft aims to provide enterprise defenders with the speed necessary to counter AI-driven cyberattacks.
Comments (0)
No comments yet. Be the first!