Microsoft launches AI cybersecurity model, agentic defense platform to cut enterprise security costs

Microsoft unveiled MAI-Cyber-1-Flash, a compact cybersecurity model built into its MDASH system for finding and fixing software vulnerabilities. The company says the model scores 96% on CyberGym, outperforming several frontier models while cutting costs by about half, and also introduced Project Perception for coordinated agentic security workflows.

Microsoft is positioning specialized, cheaper AI models as a practical alternative to relying solely on large frontier models for enterprise security. If the performance claims hold up, this could accelerate a shift toward task-specific AI systems that route work intelligently to balance accuracy, speed, and cost.