Voxtral Mini Transcribe 3
#7 in Transcription (STT)Mistral AI · v3 · mini · since 2026-10-20 · 2× · last seen Oct 22, 2026
39
Momentum
Voxtral Mini Transcribe 3 is a batch speech-to-text model from Mistral AI with 4 billion parameters, introduced as part of the Voxtral 3 release on October 20, 2026. It supports transcription, diarization and context biasing in over 50 languages and is available via the Mistral API. According to sources, the model weights are API-only.
Momentum trend
24.07.22.10.
Features
| License | Proprietary (Mistral Premier), API only |
| Platform | Mistral API (EU, US and global endpoints); public preview |
| Price | $0.0035 per minute (batch) |
| Release Date | October 20, 2026 |
| Languages | over 50 languages |