# Voxtral Small Transcribe 3 Benchmark Scores & Performance

> Voxtral Small Transcribe 3 has Mistral Voxtral 3 results results. Voice evidence is display-only and does not produce a text-model score.

Canonical page: https://benchlm.ai/models/voxtral-small-transcribe-3

Last updated: October 20, 2026

## Model Details

| Property | Value |
|----------|-------|
| Creator | Mistral AI |
| Source Type | Proprietary |
| Reasoning Type | Non-Reasoning |
| Context Window | 32K |
| Official model card | [Mistral AI model documentation](https://docs.mistral.ai/models/voxtral-small-transcribe-3-0) |
| Overall Score | Not scored (voice evidence only) |
| Overall Rank | Unranked |

## Family & Coverage

- Family: Voxtral Small Transcribe 3
- Variant: base
- Benchmarks covered: 0 of 668
- Coverage note: Speech and voice results appear below; no weighted text-model benchmark rows are stored.

## Speech and voice evidence

[Voxtral 3 launch evaluation](/voice-benchmarks/voxtral-3-launch-evaluation)

All results are Mistral-run. FLEURS macro-averages use CER for Japanese, Korean, Thai, and Chinese. Diarization averages AMI-IHM, AMI-SDM, Switchboard, Switchboard CallHome, and AISHELL-4, with 30-minute AMI and AISHELL versions. Realtime rows trade latency for accuracy; the Mini Transcribe 3 row on FLEURS 35 is an offline baseline. Voice Code Bench is a BeSimple AI test of numbers and names in support calls; the multilingual and expressive scores are judged by gemini-3.8-flash. The launch post and docs disagree on TTS language count (35+ vs 9) and realtime minimum latency (240 ms charted, 80 ms documented). Competitor rows from Mistral charts are not carried here. These results do not enter text-model or voice-agent rankings.

| Metric | Result |
| --- | --- |
| FLEURS 45-language WER · language specified | 4.0% |
| FLEURS 45-language WER · language detected | 4.1% |
| Diarization cpWER, five-set macro-average | 18.4% |
| Earnings-22 keyphrase recall · with context | 94.6% |
| Earnings-22 keyphrase recall · without context | 81.4% |

Mistral launch charts retrieved October 20, 2026. Lower error rates are better. Voice results are display-only.

- [Evaluation source](https://mistral.ai/news/voxtral-3/)

Released October 20, 2026 as a public preview in the Mistral API as `voxtral-small-transcribe-3-0`, on European, US, and global endpoints. Mistral lists 50 transcription languages, including the 24 official EU languages, plus diarization, context biasing, timestamps, and end-of-turn detection. Self-deployment is offered through Mistral sales; no weights are published.

## Other Mistral AI Models

- [Voxtral Mini Transcribe 3](/models/voxtral-mini-transcribe-3) - Score: not computed
- [Voxtral Mini Transcribe Realtime 3](/models/voxtral-mini-transcribe-realtime-3) - Score: not computed
- [Voxtral TTS 3](/models/voxtral-tts-3) - Score: not computed
- [Voxtral 4B TTS 2603](/models/voxtral-4b-tts-2603) - Score: not computed
- [Voxtral Mini 4B Realtime 2602](/models/voxtral-mini-4b-realtime-2602) - Score: not computed