At a small event in San Francisco on Monday, Microsoft unveiled its first specialized cybersecurity model, along with a new artificial intelligence cybersecurity platform, in a major blow to the major players in the space — Anthropic, Google and OpenAI.
The company describes MAI-Cyber-1-Flash as a model built for “finding difficult vulnerabilities in complex codebases”. The model was built to animate MDASH, Microsoft’s software vulnerability identification and remediation toolkit.
The new security platform is called Perception and it is designed for deploy groups of agents assisting and automating various security workflows, including bug identification and remediation. The platform can also integrate with MDASH.
The company claims that MAI-Cyber-1-Flash is significantly more powerful (and more cost-effective) than competing models based on an established AI cybersecurity benchmark.
“We are very excited to announce our results,” said Mustafa Suleyman, co-founder of DeepMind and current CEO of Microsoft AI. “We have a GPT 5.4 bound MAI-1 Cyber Flash inside the MDASH harness (sic) – it beats the Mythos 5 in Gemini, GPT 5.5 Cyber, GPT 5.6 Sol and Cyber Gym, which is the benchmark we all use. The gold benchmark.”
“We’re putting this into production immediately,” he said.
Noting that hackers are increasingly using artificial intelligence in their cyberattacks, Microsoft’s vice president of security, Hayete Gallot, described Perception as a way for enterprise defenders to “defend against AI with AI at the scale and speed that attackers have.”
A perceptual agent uses red commands, blue commands, and green commands. Red teams can provide detailed simulations of potential attacks—providing context about potential threat actors and the likely vulnerabilities they might exploit. Blue teams are dedicated to discovering and testing existing bugs, while green teams take “corrective actions” against those bugs.
Dave Weston, lead engineer at Perception, described the platform as a huge efficiency improvement for enterprise defenders. “We’ve taken hours and hours of manual work from a lot of specialized people in the security organization—appsec hunters, remediation engineers, you name it—and have it all resolved in minutes. Not only do we detect and prioritize issues, but we have detection, remediation, and even code resolution.”
While AI offers companies new defenses, its accessibility to cybercriminals has led to glaring potential threats.
The new security tools, which Microsoft says will be previewed on Nov. 3, will enter the increasingly crowded field of artificial intelligence cybersecurity solutions. Earlier this year, Anthropic Launched MythosA security platform released to a small group of partner organizations through a program called Glasswing. There is OpenAI was also launched own security solution in May through a program called Day Break.
When you purchase through links in our articles, we may earn a small commission. This does not affect our editorial independence.
