Key facts
- Microsoft launched MAI-Cyber-1-Flash, its first AI model built for cybersecurity.
- The company also released Perception, an AI platform for automating security workflows.
- MAI-Cyber-1-Flash is designed to identify vulnerabilities in complex code.
- Perception uses agentic teams to simulate attacks, detect bugs, and remediate issues.
- Microsoft claims its new model outperforms competitors on the Cyber Gym benchmark.
- The tools will be available in preview starting November 3.
Microsoft has launched its first artificial intelligence model specifically designed for cybersecurity, named MAI-Cyber-1-Flash, alongside a new AI-powered cybersecurity platform called Perception. The company announced the developments at an event in San Francisco, positioning its new tools as significant advancements in the field and direct competitors to offerings from Anthropic, Google, and OpenAI.
MAI-Cyber-1-Flash is built to identify challenging vulnerabilities within complex codebases and is integrated into Microsoft's MDASH harness for software vulnerability identification and remediation. Microsoft claims this model is more powerful and cost-effective than existing competitor models, citing performance on the established AI cybersecurity benchmark known as Cyber Gym. Mustafa Suleyman, CEO of Microsoft AI, stated that the model, integrated with GPT 5.4 within the MDASH harness, outperforms Gemini, GPT 5.5 Cyber, GPT 5.6 Sol, and Mythos 5 on this benchmark.
The Perception platform is designed to deploy teams of AI agents to automate and assist with various security workflows, including bug identification and remediation. It can also integrate with MDASH. Hayete Gallot, Microsoft's vice president for security, highlighted Perception's role in enabling enterprise defenders to counter AI-driven cyberattacks with AI at a comparable scale and speed. The platform employs agentic red teams for attack simulations, blue teams for bug detection and triage, and green teams for corrective actions.
Dave Weston, lead engineer for Perception, emphasized the platform's efficiency, noting that tasks previously requiring hours of manual work from multiple security specialists can now be accomplished in minutes, including code fixes. The introduction of these tools comes as AI is increasingly utilized by cybercriminals, creating a growing need for advanced defensive capabilities. Microsoft's new security tools are set to enter a crowded market, with Anthropic having launched its Mythos platform and OpenAI releasing its own security solution earlier this year.