Voxtral TTS 3
by Mistral
Mistral's text-to-speech model with preset voices, zero-shot cloning and inline delivery control.
Takes emotion and pace tags directly in the input, streams with low time-to-first-audio, and suits voice agents, narration and multilingual assistants. Serves through OpenRouter's speech API.
More details
Voxtral TTS 3 by Mistral is a fast-speed AI text-to-speech model. Takes emotion and pace tags directly in the input, streams with low time-to-first-audio, and suits voice agents, narration and multilingual assistants. Serves through OpenRouter's speech API. Key capabilities include 278 voices across 12 languages, emotion control, up to 4,096 characters per request.
Voxtral TTS 3 is priced at $0.02 per 1,000 characters of input text. You can generate speech with Voxtral TTS 3 on idapt.app alongside 200+ AI models in one workspace.
Voices
1 · 1 languagesAny of these voices can be paired with Voxtral TTS 3. Pick a model for quality and speed, a voice for character and language.
Provider Default
Multilingual · Neutral
Pricing
Cost per 1,000 characters
$0.02
Typical paragraph (~500 chars)
$0.008
Billed on input text length
Capabilities
Speed
Fast: Low-latency, real-time ready
Emotion Control
Steer happy / sad / angry / neutral delivery
Max Input
4,096 characters per request
Voices
1 voices · 1 languages