← All models

Voxtral Mini Transcribe 3

Audio Preview Full comparison ↗

mistral/voxtral-mini-transcribe-3 · by Mistral AI · speech-transcription-model

Pricing — 1 offering(s)

Mistral AI

About this offering: Mistral's small speech-to-text model (4B parameters), in Public Preview since 2026-10-20, at $0.0035 per minute of audio. It transcribes 50 languages with speaker diarization, timestamps and context biasing (key terms to improve recall), per Mistral. Mistral also lists a cached-input rate of $0.00035 per minute, which isn't tracked as a separate price here.

Audio minutes

  • $0.0035 / minute Current 2026-10-20 → present

Showing the active price and any recorded history. Full pricing history is available via the paid API — see API docs.

See how Voxtral Mini Transcribe 3 fits into a cost-aware routing setup

See how →

Capability profile

How Voxtral Mini Transcribe 3 rates across core capability dimensions, with the task-level evidence behind each rating.

transcription accuracy Strong

Mistral reports a 4.6% macro-average WER on 45 FLEURS languages (language specified), against 4.1% for Voxtral Small Transcribe 3 and 16.0% for Whisper large-v3 in the same chart. Vendor-reported.

language support Strong

50 languages, including the 24 official EU languages, per Mistral's launch post; vendor-reported.

speaker diarisation Strong

Native diarization; Mistral reports 18.6% cpWER over AMI, Switchboard, CallHome and AISHELL-4 (Small: 18.4%). Vendor-reported.

real time streaming Weak

Offline (file) transcription; Mistral's realtime model is the separate Voxtral Mini Transcribe Realtime 3.

custom vocabulary Strong

Context biasing: key terms and free-form text bias the transcription, per Mistral.

Benchmarks

Benchmark Score Config Source
FLEURS 45-language WER (language specified) 4.6 % Mistral's launch post: macro-average WER (CER for JA/KO/TH/ZH) over the 45 FLEURS languages all seven compared models support; 4.5% with the language auto-detected. Vendor-reported. source ↗
Diarization cpWER 18.6 % Mistral's launch post: macro-average concatenated minimum-permutation WER over AMI-IHM, AMI-SDM, Switchboard, CallHome and AISHELL-4 (30-minute AMI and AISHELL versions). Vendor-reported. source ↗

Operator guidance

Choose Voxtral Mini Transcribe 3 for cheap batch transcription that needs diarization and custom vocabulary across many languages. Universal-2 is cheaper and GA; Voxtral Small Transcribe 3 costs a little more for Mistral's best accuracy. Its accuracy figures are vendor-reported.

Use cases

Limitations

Citations