Microsoft releases its first cybersecurity AI model, MAI-Cyber-1-Flash, and launches the Perception security platform. Microsoft officially launched its first cybersecurity professional model, MAI-Cyber-1-Flash, and its new AI cybersecurity platform, Perception, at a recent event in San Francisco.

Microsoft releases its first cybersecurity AI model, MAI-Cyber-1-Flash, and launches the Perception security platform. Microsoft recently officially launched its first cybersecurity professional model, MAI-Cyber-1-Flash, and its new AI cybersecurity platform, Perception, at an event held in San Francisco recently, further increasing its focus on the field of artificial intelligence security defense and competing with companies such as Anthropic, Google, and OpenAI. Microsoft said that MAI-Cyber-1-Flash is a model specially built for network security scenarios. The goal is to find difficult vulnerabilities in complex code bases and provide core capability support for Microsoft's vulnerability identification and repair tool MDASH. Combined with the Perception platform, this model can help enterprises automate security processes such as vulnerability discovery, analysis, and repair. Microsoft AI CEO Mustafa Suleyman said that the company has combined MAI-1 Cyber ​​Flash with GPT5.4 in the MDASH framework and surpassed competing models such as Gemini, GPT5.5Cyber, GPT5.6Sol and Mythos5 in the Cyber ​​Gym network security benchmark test. Microsoft says the model offers advantages in performance and cost efficiency and plans to put it into production immediately. At the same time, the Perception platform launched by Microsoft adopts a multi-agent collaboration architecture and consists of three types of AI agents: red team, blue team and green team. Among them, the red team agent is responsible for simulating potential attack behaviors and analyzing threat paths; the blue team agent is used to detect and classify vulnerabilities; and the green team agent performs vulnerability repair and security optimization operations. Hayete Gallot, vice president of security at Microsoft, said that as hackers increasingly use AI to launch attacks, companies need to use AI to improve defense efficiency. Perception is designed to help security teams leverage AI to protect at the speed and scale of attackers. Dave Weston, chief engineer at Perception, said that in the past, security teams required multiple professionals to spend hours to complete vulnerability troubleshooting and repair. With the help of AI agent collaboration, enterprises can complete vulnerability discovery, prioritization, security detection and code repair in minutes. Microsoft plans to launch a preview version of the security tool on November 3. With Anthropic previously launching the Mythos security platform and opening it to some partners through the Glasswing project, and OpenAI launching the Daybreak security solution in May this year, competition in the field of AI network security is further heating up. via AI new