Primary use stated or inferred cautiously from official documentation.
M
MISTRAL / voxtral-mini-tts-3-0
Voxtral TTS 3
Voice synthesis model presented on October 20, 2026. The current specific guide declares 35 languages and more than 50 dialects; the old count of nine languages from the model card is not used as the current limit. It converts text to audio, with an optional voice reference. Batch and streaming responses use the same model; they are not duplicated as distinct checkpoints.
01 / SUMMARY
The essential
Maximum input capacity when published by the source.
Documented maximum generation limit.
Verified availability channels.
Primary functional family and verified specialties.
Declared terms for API access, model weights, or self-hosting.
02 / TECHNICAL SHEET
Limits and integration
| API ID | voxtral-mini-tts-3-0 |
|---|---|
| Model type | Voice and audio |
| Access model | Paid proprietary |
| License | Proprietary |
| Deployment | Hosted API |
| Release | 20 OCT 2026 |
| Knowledge cutoff | Not published |
| Entrance | Text · Audio |
| Exit | Audio |
| Context window | Text and optional voice reference |
| maximum output | Synthesized audio |
| Reasoning | Not published |
| Published tools | |
| Structured Outings | Not published |
| Batch processing | Not published |
| Prompt cache | Not published |
| Fine-tuning | Not published |
| Verified platforms | Mistral API |
Narration, accessibility, and spoken interfaces via /v1/audio/speech. It supports emotion tags and slow or fast speeds. Use voice references only with verifiable authorization and check pronunciation, numbers, and language changes before publishing. Multilingual support does not guarantee identical quality in all accents nor eliminate human review of sensitive messages.
The official migration indicates that free-form style instructions are not available at launch: we do not present them as a usable feature. Emotion tags are a different mechanism. Do not use cloning to impersonate people nor interpret a voice reference as consent. Preserve the versioned identifier and validate format, duration, and audio processing conditions.
03 / CAPABILITIES
What it can do
Orientation
Narration, accessibility, and spoken interfaces via /v1/audio/speech. It supports emotion tags and slow or fast speeds. Use voice references only with verifiable authorization and check pronunciation, numbers, and language changes before publishing. Multilingual support does not guarantee identical quality in all accents nor eliminate human review of sensitive messages.
Context
Text and optional voice reference · Synthesized audio
Tools and integration
Not published
Access
Mistral API
Input modalities
TextAudio
04 / PRICES
Documented cost
| Concept | Worth | Unit/condition |
|---|---|---|
| Standard input | 13 USD / million characters | Standard API rate |
| Cached input | Not published | Reading reused prefixes |
| Cache write or storage | Not published | The condition varies by provider |
| Standard output | Included in the text rate | May include reasoning tokens |
| Batch input | Not published | Asynchronous processing |
| Batch output | Not published | Asynchronous processing |
Public rates in dollars per minute of audio or per million characters, depending on the variant; they are not token prices. Numerical fields per million tokens are left blank to avoid false comparisons. Confirm billing conditions and availability before integrating. Private enterprise deployment requires a specific agreement; no downloadable weights have been published for these variants.
i Prices change and may depend on level, region or context length. Check the source before making a decision.
05 / EVALUATIONS
How to read the results
“
Mistral uses VoiceCodeBench with 300 English prompts and two voices; the multilingual test covers 377 prompts in eleven languages. These are provider evaluations with an automatic judge, not Inferama's own tests nor a measure of clinical quality. We do not convert a judge's preferences into factual accuracy. Review intelligibility, naturalness, pronunciation, and authorized use of the voices.
i Provider evaluation; numerical scores not verified by Inferama.
Inferama only highlights a “best result” when the metric, test set, configuration, and date allow for an equivalent comparison. The supplier's figures are presented as claims from its own source.
06 / VERSIONS
Chronology
Voxtral TTS 3
Version added to Inferama's verified catalog.
ANALYSIS
Related analysis
07 / SOURCES