Voxtral TTS

About

Voxtral TTS is Mistral AI's first text-to-speech model, released March 26, 2026. It is a 4B parameter open-weight model that generates human-quality speech from text across nine languages. The model architecture splits into three components: a 3.4B transformer decoder backbone, a 390M acoustic transformer, and a 300M neural audio codec. What makes Voxtral stand out is the combination of quality and accessibility. Human evaluations show it produces more natural-sounding speech than ElevenLabs Flash v2.5, and matches ElevenLabs v3 quality — while the weights are freely available on HuggingFace. You can run it on a consumer GPU or laptop, or use Mistral's API at $0.016 per 1,000 characters. Voice cloning requires just 3 seconds of reference audio. The model captures accent, inflections, intonations, and even speech disfluencies from that tiny sample, then applies them across any of the nine supported languages. Zero-shot cross-lingual adaptation means you can clone an English voice and have it speak fluent French with the original speaker's characteristics preserved. Latency is 70ms for typical inputs (500 characters generating a 10-second clip), with a real-time factor of approximately 9.7x. The API handles arbitrarily long content through smart interleaving, and the model natively generates up to 2 minutes of audio per request. The open-weight release under CC BY NC 4.0 means researchers and hobbyists can run it locally, fine-tune it, and integrate it into non-commercial projects. For commercial use, the API is the intended path at $0.016 per 1K characters — roughly 10x cheaper than ElevenLabs' standard pricing tier. For developers building voice applications, accessibility tools, or content creation pipelines, Voxtral is the first serious open-weight alternative to proprietary TTS services. The quality-to-cost ratio is unprecedented in the TTS market.

Key Features

• 4B parameter open-weight model with 3.4B transformer decoder, 390M acoustic transformer, and 300M audio codec
• 9 languages: English, French, German, Spanish, Dutch, Portuguese, Italian, Hindi, Arabic
• Voice cloning from just 3 seconds of reference audio with accent and inflection preservation
• 70ms model latency for typical 500-character inputs generating 10-second audio clips
• 9.7x real-time factor — generates audio nearly 10x faster than playback speed
• Zero-shot cross-lingual voice adaptation (clone English voice, generate French speech)
• Emotion steering support for expressive, context-aware speech generation
• Native generation of up to 2 minutes per request, API handles arbitrary length via smart interleaving
• Runs on consumer hardware: modern laptops, mid-range desktop GPUs, some high-end mobile devices
• Open weights on HuggingFace (mistralai/Voxtral-4B-TTS-2603) for local deployment

Use Cases

• 1AI voice assistants and conversational interfaces with natural, low-latency responses
• 2Podcast and audiobook production with voice cloning from minimal reference audio
• 3Accessibility tools: screen readers, document narration, and real-time text-to-speech for visually impaired users
• 4Multilingual content localization — generate voiceovers in 9 languages with consistent voice identity
• 5Educational content creation: lecture narration, language learning apps, interactive tutorials
• 6Customer support IVR systems with natural-sounding, branded voice responses
• 7Game development: NPC dialogue generation with diverse voice characteristics
• 8Content creator workflows: YouTube narration, TikTok voiceovers, social media audio