答案已发布25 来源
微软自研AI模型再出重拳：图像生成和语音模型GPU成本直降89%
微软在2026年7月23 24日推出两全新自研AI模型：MAI Image 2.5 Pro（最高画质图像生成）与MAI Voice 2 Flash（低延迟语音合成）[1][3][4]。 已部署至Bing Image Creator（完全自研化）、PowerPoint、OneDrive及Dynamics 365 Contact Center，其中T Mobile和EasyJet等客户已接入语音模型[2][3][4]。
AI 提示
openai.comCreate a landscape editorial hero image for this Studio Global article: Search & fact-check with cited sources for What two new in-house AI models did Microsoft launch on Thursday, how are they being deployed acr. Article summary: Here is a comprehensive answer based on the strongest available sources.. Topic tags: general, general web, user generated. Style: premium digital editorial illustration, source-backed research mood, clean composition, high detail, modern web publication hero. Use reference image context only for broad subject, composition, and topical grounding; do not copy the exact image. Avoid: logos, brand marks, copyrighted characters, real person likenesses, fake screenshots, UI text, readable text, watermarks, charts with fake numbers, clickbait thumbnails, icons, and tiny thumbnail layouts. Make it useful as an illustrative visual, not as factual evidence.
两款新模型解读
- MAI-Image-2.5-Pro：微软迄今最高保真度的图像生成模型，专为专业级视觉素材、精细编辑和在生成图像中准确呈现文字而打造
。它通过Microsoft Foundry定价，每百万文本输入tokens收费5美元，每百万图像输入tokens收费8美元，每百万图像输出tokens收费106美元
。
- MAI-Voice-2-Flash：一款低延迟文本转语音模型，针对速度和规模化进行了优化，适用于客服和实时语音场景
。微软称其速度是标准版MAI-Voice-2的两倍，性价比高出32%
。
Studio Global AI
Search, cite, and publish your own answer
Use this topic as a starting point for a fresh source-backed answer, then compare citations before you share it.
人们还问
“微软自研AI模型再出重拳：图像生成和语音模型GPU成本直降89%”的简短答案是什么？
微软在2026年7月23 24日推出两全新自研AI模型：MAI Image 2.5 Pro（最高画质图像生成）与MAI Voice 2 Flash（低延迟语音合成）[1][3][4]。
首先要验证的关键点是什么？
微软在2026年7月23 24日推出两全新自研AI模型：MAI Image 2.5 Pro（最高画质图像生成）与MAI Voice 2 Flash（低延迟语音合成）[1][3][4]。 已部署至Bing Image Creator（完全自研化）、PowerPoint、OneDrive及Dynamics 365 Contact Center，其中T Mobile和EasyJet等客户已接入语音模型[2][3][4]。
接下来在实践中我应该做什么？
GPU推理基础设施成本对比OpenAI同类模型：PowerPoint中MAI Image 2.5 Pro最高降低84%，Dynamics 365中MAI Voice 2 Flash最高降低89%[3][4]。
来源
- thurrott.comMicrosoft AI Announces New Model Variants, Minor Shift ...
- finance.biggo.comMicrosoft Unveils In-House Image and Voice AI Models to Reduce OpenAI Dependency — BigGo Finance
- microsoft.aiBringing Ode Poetry to life with MAI's audio models - Microsoft AI
- microsoft.aiIntroducing MAI-Image-2.5-Pro and MAI-Voice-2-Flash
- kucoin.comMicrosoft launches in-house AI models, reducing GPU ...
- x.comMicrosoft has launched two new in-house AI models ...