Microsoft CEO Nadella stated that its self-developed MAI series models perform on par with GPT-5.6 in most common Excel tasks, and because they do not rely...
行业动态IT之家(RSS)
Today AI Intelligence Brief
Microsoft CEO Nadella stated that its self-developed MAI series models perform on par with GPT-5.6
in most common Excel tasks, and because they do not rely on the latest AI accelerators, they can run on NVIDIA H100/A100 GPUs, making deployment costs lower. The MAI models have achieved significant progress on GitHub Copilot and Excel through reinforcement learning environments (RLEs) targeted at specific products. Nadella emphasized that in the future, enterprises will orchestrate models based on tasks rather than relying entirely on a single cutting-edge model.
Original Article Excerpt
IT之家:https://www.ithome.com/ 7 月 25 日消息,科技媒体 firstpost 于 7 月 23 日发布博文,报道称微软首席执行官萨提亚 · 纳德拉(Satya Nadella)表示, 微软自研 MAI 系列模型在许多应用场景中超越了前沿 AI,而且能降低企业部署 AI 成本。
纳德拉表示微软正在推进“前沿扩散与控制”(Frontier Diffusion & Control)战略,认为在未来企业 AI 场景下,客户不会再完全依赖单一前沿模型,会根据实际任务和需求,编排最合理的 AI 模型调用系统。
The readable text on this page was extracted from the public source and organized with attribution, publication time and the original link. Copyright remains with the original author and publisher.
Microsoft CEO Nadella stated that its self-developed MAI series models perform on par with GPT-5.6 in most common Excel tasks, and because they do not rely...
IT之家(RSS)2026-07-25T07:39:52.000Z
Scan to open this article
Aioga aggregates global AI updates and preserves source information for verification and citation.