Microsoft has unveiled its first artificial intelligence model designed specifically for cybersecurity, alongside a new platform powered by AI agents to automate cyber defence operations. The launch marks the company’s latest effort to strengthen organisations’ ability to detect and respond to increasingly sophisticated cyber threats.
The company introduced MAI-Cyber-1-Flash, an AI model developed to identify complex software vulnerabilities. The model is integrated with the MDASH system, which analyses security flaws and assists in their remediation, helping organisations accelerate vulnerability management and improve overall cyber resilience.
Microsoft also announced Perception, a new platform that employs teams of AI agents to detect and analyse vulnerabilities, recommend mitigation strategies, and support security teams throughout the remediation process.
According to Microsoft, MAI-Cyber-1-Flash delivered advanced performance in the Cyber Gym benchmark, which evaluates the capabilities of artificial intelligence models in cybersecurity. The company said the platform includes specialised AI teams capable of simulating cyberattacks, identifying vulnerabilities, and carrying out corrective actions to strengthen digital defences.
The launch comes amid intensifying competition among major technology companies to develop AI-powered cybersecurity solutions, as organisations seek more effective ways to counter increasingly complex and rapidly evolving cyber threats.
Microsoft said the new tools will be made available in preview starting 3 November.
