微软发布自有网络安全模型 MAI-Cyber-1-Flash，最棘手任务仍依赖 OpenAI

微软表示，它在 AI 网络安全方面已经缩小了与前沿模型的差距。该公司正在推出 MAI-Cyber-1-Flash：https://microsoft.ai/news/introducing-mai-cyber-1-flash-inside-mdash/，这是一个内置于此前推出的 MDASH 多代理系统中的紧凑型安全模型。该组合在 CyberGym（一个衡量 AI 在大型代码库中发现真实安全漏洞能力的基准测试）中得分为 96%，比 Mythos 高出 12 分，并领先于 Gemini 和 GPT。微软表示，由于 MAI-Cyber-1-Flash 可以处理 90% 的任务，仅将复杂案例交给 GPT-5.4，因此成本预计将下降 50%。该模型基于 MAI-Thinking-1 系列：https://the-decoder.com/microsoft-trained-its-mai-models-on-unlicensed-web-data-despite-promising-enterprise-grade-clean-and-commercially-licensed-data/。

微软仍然依赖 OpenAI 来处理复杂推理，但这一安排符合其作为 AI 模型协调者的不断发展角色：https://the-decoder.com/microsoft-ceo-satya-nadella-warns-of-a-small-number-of-ai-systems-capturing-all-the-economic-returns/。同样的转变也使这家雷德蒙德公司成为开放权重的倡导者：https://the-decoder.com/microsofts-open-weight-ai-push-is-so-obviously-an-azure-play-it-hurts/，此前多年通过独家 OpenAI 分发推动 Azure 增长。

微软还推出了 Perception，这是一种基于代理的安全系统，可实时监控和缓解威胁，利用其每天超过 100 万亿条安全信号和 160 万客户的数据优势。广告 DEC_D_Incontent-1 广告

保持对 AI 的关注。清晰、有用、不空洞。

关注 The Decoder 获取 AI 新闻、背景故事和专家分析。

The Decoder：https://the-decoder.com/

Microsoft says it has closed the gap with frontier models in AI cybersecurity. The company is introducing MAI-Cyber-1-Flash：https://microsoft.ai/news/introducing-mai-cyber-1-flash-inside-mdash/, a compact security model built into its previously unveiled MDASH multi-agent system. The combination scores 96 percent on CyberGym, a benchmark measuring how well AI spots real security flaws in large codebases, putting it 12 points above Mythos and ahead of both Gemini and GPT. Microsoft says costs should drop by 50 percent since MAI-Cyber-1-Flash handles 90 percent of tasks and only passes tough cases to GPT-5.4. The model is based on the MAI-Thinking-1 line：https://the-decoder.com/microsoft-trained-its-mai-models-on-unlicensed-web-data-despite-promising-enterprise-grade-clean-and-commercially-licensed-data/.

Microsoft still relies on OpenAI for complex reasoning, but the setup fits its evolving role as an AI model orchestrator：https://the-decoder.com/microsoft-ceo-satya-nadella-warns-of-a-small-number-of-ai-systems-capturing-all-the-economic-returns/. That same shift has turned the Redmond company into an open-weights advocate：https://the-decoder.com/microsofts-open-weight-ai-push-is-so-obviously-an-azure-play-it-hurts/ after years of fueling Azure growth through exclusive OpenAI distribution.

Microsoft is also launching Perception, an agent-based security system that monitors and mitigates threats in real time, citing its data advantage of over 100 trillion daily security signals and 1.6 million customers. Ad DEC_D_Incontent-1 Ad

Stay in the loop on AI. Clear, useful, no fluff.

Follow The Decoder for AI news, background stories and expert analyses.

The Decoder：https://the-decoder.com/