Transcription with Voxtral 3 models starts at 0.0035 USD per minute, SDK uses LiveKit
Transcription pricing in the Voxtral 3 model family starts at 0.0035 USD per minute. The SDK for voice agents uses LiveKit, and according to Mistral AI, speech synthesis adapts emotion, pace and emphasis to the context of the conversation.
Transcription in the Voxtral 3 model family starts at 0.0035 USD per minute; Mistral AI assigns this price to the Mini variant with 4 billion parameters. The report further specifies that the SDK for voice agents is based on LiveKit. According to the company, speech synthesis adapts emotion, pace and emphasis to the context of the conversation.
Mistral AI released Voxtral Small Transcribe 3 for transcription and Voxtral Mini TTS 3 for speech synthesis, along with an SDK that connects listening, reasoning, tool calling and speaking in real time. For transcription, the company alternately states 50 and more than 50 languages. According to the company, speech synthesis is expected to support over 35 languages; sources give the number of dialects as both 50 and more than 50. The company quotes a speech synthesis price of 13 USD per million characters.
According to the company, the model family is available through an API hosted in the EU and for deployment on an organization's own infrastructure. The company says this allows organizations to decide where audio is processed and whether conversations are retained.
Why it matters
A voice agent can work with connected systems during a phone or web conversation, for example to book an appointment or check an invoice. The quoted per-minute price applies to transcription; on its own, it does not determine the cost of an entire conversation, which also includes reasoning, tool calling and speech synthesis.
What was added since the original report
Verified updates
-
Transcription pricing starts at 0.0035 USD per minute.; The SDK is based on LiveKit.; Speech synthesis supports emotional adaptation.
- Transcription pricing starts at 0.0035 USD per minute.
- The SDK is based on LiveKit.
- Speech synthesis supports emotional adaptation.
Two audiences, two different impacts
What this means
For individuals
A developer can use the SDK to connect a voice interface to their own tools and build applications that also perform tasks during a conversation.
For a business
When designing voice-based customer service, a company gains a choice between an API hosted in the EU and its own infrastructure, which affects where audio is processed and the rules for retaining conversations.
ProcessesCheck the original
Event sources
clearly official source · 1 publisher, 0 independent. We count feeds from the same owner only once.