<p id="speakable-summary" class="wp-block-paragraph">Microsoft on Monday launched its first cybersecurity-specialized model alongside a new AI cybersecurity platform at a small event in San Francisco, taking a big swipe at major players in the space — namely Anthropic, Google, and OpenAI.</p>
<p class="wp-block-paragraph">The company describes <a href="https://microsoft.ai/news/introducing-mai-cyber-1-flash-inside-mdash/" target="_blank" rel="noreferrer noopener nofollow">MAI-Cyber-1-Flash</a> as a model that’s built “to find challenging vulnerabilities in complex codebases.” The model is built to animate MDASH, Microsoft’s harness dedicated to software vulnerability identification and remediation.</p>
<p class="wp-block-paragraph">The new security platform is dubbed Perception, and it’s designed to <a href="https://blogs.microsoft.com/blog/2026/07/27/rethinking-security-for-the-age-of-ai/" target="_blank" rel="noreferrer noopener nofollow">deploy teams of agents</a> to assist with and automate various security workflows, including identifying and remediating bugs. The platform can also integrate with MDASH.</p>
<p class="wp-block-paragraph">The company claims MAI-Cyber-1-Flash is significantly more powerful (and more cost-effective) than competitor models, based on its performance on an established AI cybersecurity benchmark. </p>
<p class="wp-block-paragraph">“We’re very very excited to announce our results,” said Mustafa Suleyman, the co-founder of DeepMind and current CEO of Microsoft AI. “We have MAI-1 Cyber Flash binded [sic] with GPT 5.4 inside of the MDASH harness — which beats out Gemini, GPT 5.5 Cyber, GPT 5.6 Sol, and Mythos 5 on Cyber Gym, which is the primary benchmark that we all use. The golden benchmark.”</p>
<p class="wp-block-paragraph">“We’re shipping this into production immediately,” he added.</p>
<p class="wp-block-paragraph">Noting that hackers are increasingly using AI in their cyberattacks, Hayete Gallot, Microsoft’s vice president for security, described Perception as a way for enterprise defenders to “defend against AI with AI at the scale and speed that the attackers have.”</p>
<p class="wp-block-paragraph">Perception uses agentic red teams, blue teams, and green teams. The red teams can provide detailed simulations of potential attacks — providing context about potential threat actors and the likely vulnerabilities that they might exploit. Blue teams are dedicated to detecting and triaging existing bugs, while green teams take “corrective actions” against those bugs.</p>
<p class="wp-block-paragraph">Dave Weston, the lead engineer for Perception, described the platform as a massive efficiency upgrade for corporate defenders. “We’ve gone from this taking hours and hours of manual work from multiple specialized folks across the security organization — appsec hunters, remediation engineers, you name it — and in minutes, we have a fix for all of this. Not only do we discover the issues and prioritize them, but we have detection, posture fixing, and even a code fix.”</p>
<p class="wp-block-paragraph">Though AI has offered new defensive capabilities to companies, its availability to cybercriminals has given rise to a dazzling array of potential threats.</p>
<p class="wp-block-paragraph">Microsoft’s new security tools, which the company said will be available in preview on November 3, will enter an increasingly crowded field of AI cybersecurity solutions. Earlier this year, Anthropic <a href="https://techcrunch.com/2026/04/07/anthropic-mythos-ai-model-preview-security/?_thumbnail_id=3085495" target="_blank" rel="noreferrer noopener">launched Mythos</a>, a security platform that was released to a small coterie of partner organizations through a program called Glasswing. OpenAI has <a href="https://www.theverge.com/ai-artificial-intelligence/928342/openai-daybreak-security-ai" target="_blank" rel="noreferrer noopener nofollow">also launched</a> its own security solution in May through a program called Daybreak.</p>
© 2026 Winzheng.com 赢政天下 | 转载请注明来源并附原文链接