Mistral AI·· 2026-03-24
Speaking of Voxtral
Speaking of Voxtral
AI summary
Voxtral TTS: A frontier, open-weights text-to-speech model that’s fast, instantly adaptable, and produces lifelike speech for voice agents.
Selection record
Threshold 60Official, first-handFirst 71Second 72
AdmittedSum of both 143 ≥ twice the threshold 120
- Source tier
- Official, first-hand; this tier's threshold is 60
- Pre-filter
- passed:发布Voxtral TTS语音模型,属AI技术
- Why it was chosen
- The 4B parameter count, 70ms latency, nine-language support and comparison with ElevenLabs help assess voice-agent cost and quality trade-offs.
A model scores each item twice, independently, against one written standard, out of 100. An item is admitted only when the two scores add up to twice the threshold. The threshold is set per source tier.
Source: Mistral AI · mistral.ai