Microsoft Launches First Cybersecurity AI Model and New Agentic Defense Platform
Microsoft
Microsoft unveiled MAI-Cyber-1-Flash, its first AI model specialized for cybersecurity, along with a new agentic platform called Perception at an event in San Francisco. The model is designed to find vulnerabilities in complex codebases and is paired with the MDASH harness. Perception deploys teams of AI agents (red, blue, green) to automate security workflows.
Microsoft announced its first cybersecurity-specialized AI model, MAI-Cyber-1-Flash, at a small event in San Francisco, alongside a new AI security platform called Perception. The model is built to find challenging vulnerabilities in complex codebases and operates within Microsoft's MDASH harness for software vulnerability identification and remediation. According to Mustafa Suleyman, CEO of Microsoft AI, the model outperforms competitors like Gemini, GPT-5.5 Cyber, GPT-5.6 Sol, and Mythos 5 on the Cyber Gym benchmark. Perception is an agentic platform that uses red, blue, and green teams to simulate attacks, detect bugs, and apply fixes automatically. The platform aims to speed up security workflows from hours to minutes. Microsoft will release previews on November 3.
Source: TechCrunch AI —
original