微软AI发布面向语音代理的转录与文本转语音新模型
原标题：Microsoft AI releases new transcription and text-to-speech models for voice agents
AI 摘要
微软AI发布实时转录模型MAI-Transcribe-2-Streaming，声称在Artificial Analysis准确率排名第一，支持60种语言，首个部分结果延迟约100毫秒，可让语音代理在用户说话中途响应，年底前音频价格为每小时0.54美元。同时发布两款文本转语音模型MAI-Voice-2.1和MAI-Voice-2.1-Flash，前者可用同一声音说23种语言并带本地口音，后者延迟150毫秒、每百万字符15美元。两款语音模型均可从几秒参考音频克隆声音，并内置防滥用保护，已上线Microsoft Foundry、MAI Playground和OpenRouter。
正文节选
Microsoft AI releases new transcription and text-to-speech models for voice agents Microsoft AI has released MAI-Transcribe-2-Streaming, a new model for real-time transcription. Microsoft says it ranks first for accuracy on Artificial Analysis. The model transcribes 60 languages and delivers its first partial results in just over 100 milliseconds. Microsoft says this lets voice agents respond while someone is still mid-sentence. Through the end of the year, an hour of audio costs $0.54 at the in