Mistral launches Voxtral 3 voice model series and voice agent SDK
BIBIBI
AT A GLANCE
Mistral launches the Voxtral 3 voice model series and a LiveKit-based voice agent SDK。
Article
Mistral launches Voxtral 3 voice model series and introduces a voice agent SDK
PANews October 21 News, according to an official announcement from Mistral, the French AI company Mistral has launched the Voxtral 3 series of voice models, including the speech-to-text model Voxtral Small Transcribe 3、Voxtral Mini Transcribe 3, the real-time transcription model Voxtral Mini Transcribe Realtime 3 and the text-to-speech model Voxtral Speech 3, and introduced a LiveKit-based voice agent SDK, supporting speaker diarization, voice cloning, function calling, and SIP telephone integration. Its transcription models support 50 multiple languages,API priced per minute 0.0035 to 0.008 USD, while text-to-speech is priced per 1000000 character 13 USD. Mistral said,...
PANews said the news came from an official Mistral announcement.
02
Mistral launches the Voxtral 3 series of voice models.
03
The series includes Voxtral Small Transcribe 3、Voxtral Mini Transcribe 3、Voxtral Mini Transcribe Realtime 3 and Voxtral Speech 3。
04
Mistral introduces a LiveKit-based voice agent SDK, supporting speaker diarization, voice cloning, function calling, and SIP telephone integration.
05
The transcription models support 50 multiple languages,API priced per minute 0.0035 to 0.008 USD; text-to-speech is priced per 1000000 character 13 USD。
AI-assisted interpretation
The following is analysis, separate from reported facts. Verify important claims independently.
This product suite covers speech-to-text, real-time transcription, and text-to-speech, and provides a software development kit for building voice agents.
Why it matters to readers
Developers can handle multilingual transcription, speech generation, telephone integration, and other voice-agent functions within the same product series.
Beginners can understand it as a set of development tools that enable applications to understand speech, record conversations in real time, generate voices, and make and receive phone calls.
Risks and unknowns
The provided text is truncated at “Mistral said,” so subsequent statements cannot be confirmed.
The text does not provide Mistral’s original official announcement, so the model capabilities and pricing details cannot be fully verified on this basis.
The text does not specify the actual availability date, regions, or usage restrictions for the models and SDK.
Related Developments
Loading event timeline…
Related concepts
Speech-to-Text
This term is not in the glossary yet. Browse related concepts in the glossary.
