Technology
Mistral's free text-to-speech AI, explained
A high-quality AI voice you can run on your own computer. Here's what that means.
Mistral's Voxtral (March 2026) is a free-to-download AI voice model that runs on one GPU.
AI that reads text aloud in a natural voice has become very good — and Mistral's Voxtral is notable because you can run it yourself, for free. Here's the plain-English version.
What makes it special
Most high-quality AI voices are 'rented' — you use a company's website or service and pay per use. Voxtral is small enough to run on a single good graphics card, supports nine languages, and in Mistral's own tests sounded as good as (or better than) a leading paid service called ElevenLabs. That makes it useful if you care about cost, privacy, or running things offline.
The free-versus-paid catch
It's free for personal and non-commercial use — hobby projects, learning, experiments. If you want to use it in a business, you go through Mistral's paid online service instead (still much cheaper than many rivals). So 'free' has an asterisk depending on what you're doing.
Sources
- Mistral AI just released a text-to-speech model it says beats ElevenLabs — and it's giving away the weights for free — VentureBeat, 26 March 2026
- Mistral releases an open-weights 'speaking' AI model with Voxtral TTS — SiliconANGLE, 26 March 2026