The SDK combines listening, reasoning, tool calling, and speaking in a real-time streaming pipeline. The company says Voxtral Small Transcribe 3 supports transcription in 50 languages, with diarization, context biasing, and end-of-turn detection; Voxtral Mini TTS 3 supports 35+ languages and 50+ dialects. Mistral also describes Voxtral Realtime Transcribe 3 as delivering sub-500ms latency in 35+ languages. Voxtral 3 is available through Mistral’s EU-hosted API or for self-deployment. Mistral lists Mini Transcribe 3 at $0.0035 per minute and TTS at $13 per million characters. These performance comparisons and product claims are from Mistral’s announcement.
Voxtral 3: Voice Agent SDK & Speech Models in 50+ Languages
In one minute
Mistral announced Voxtral 3, a family of speech models and a voice-agent SDK.
What changed: Mistral has moved from reported voice-mode development to announcing Voxtral 3, a voice-agent SDK and speech-model family for 50+ languages.