● Covered by 1 source · 1 reportMedium impact1 neutral
Microsoft Introduces New Voice AI Models for Real-time Speech Processing
Key points
- Microsoft Foundry released three new voice AI models.
- MAI-Transcribe-2-Streaming transcribes live speech into text.
- MAI-Voice-2.1 and MAI-Voice-2.1-Flash are text-to-speech models.
- Models use "partials" for incremental speech recognition.
New AI Models for Voice Interaction
Microsoft Foundry, Redmond's app and agent model catalog, has introduced three new Microsoft AI (MAI) models. These models are designed to improve the responsiveness and efficiency of voice-based AI applications.
The new additions include MAI-Transcribe-2-Streaming, which converts live speech into text as it is spoken, and two text-to-speech models: MAI-Voice-2.1 for expressive multilingual speech, and MAI-Voice-2.1-Flash, optimized for high-volume, responsive voice applications.
Incremental Speech Recognition
A key feature of the new transcription technology is its use of "partials," which are initial hypotheses generated within milliseconds of receiving audio. This incremental speech recognition allows the system to begin processing a request before a speaker has finished their utterance.
The model refines these partials as more context becomes available, committing to a stable transcript once the utterance is complete. This approach reduces delays and enables quicker responses from voice agents and assistants.
Applications and Benefits
The ability to transcribe and process speech in real-time allows applications to act more quickly. For instance, a customer service agent can start identifying a caller's request before a sentence is finished, and voice assistants can begin reasoning or preparing tool calls sooner.
MAI-Voice-2.1-Flash is suitable for live support agents and voice assistants requiring responsive, high-volume interactions. MAI-Voice-2.1 is recommended when expressive fidelity in speech generation is the primary concern.
✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →
The daily brief
One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.
One email a day. Unsubscribe in one click, any time.
Today's brief
Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.
▶ Play today's brief
New every morning, and the back catalogue is archived by date.
Reporting from
Microsoft Foundry has released three new AI models, including MAI-Transcribe-2-Streaming, MAI-Voice-2.1, and MAI-Voice-2.1-Flash. These models enable real-time speech-to-text transcription and expressive text-to-speech, allowing applications to process and respond to spoken requests more quickly.