TREE NEWS update: French AI company Mistral released the Voxtral 3 family of speech models, including the Voxtral Small Transcribe 3, Voxtral Mini Transcribe 3, Voxtral Mini Transcribe Realtime 3 transcription models and the Voxtral Speech 3 text-to-speech model, along with a voice agent SDK built on LiveKit that supports speaker diarization, voice cloning, function calling and SIP telephony. The transcription models support more than 50 languages, with API pricing from $0.0035 to $0.008 per minute and text-to-speech at $13 per million characters.
Mistral launches Voxtral 3 speech models and voice agent SDK
AI take
The significance here is less any single model than the bundling: a full speech stack — transcription, TTS and an agent SDK with telephony — from a European vendor, priced per minute and per character. That puts Mistral in direct competition with established speech API providers and with the telephony plumbing that voice-agent startups currently assemble themselves. The SIP support and diarization are the practically load-bearing features for enterprise call handling, not the language count. Whether the LiveKit-based SDK attracts developers away from incumbent stacks, and how the per-minute pricing holds up at volume, is the open question worth watching.
Generated by AI for reference only.
Share on WeChat
Open WeChat → Scan → then tap "…" to send to a chat or Moments.
Tap "…" in the top-right corner to send to a chat or share to Moments.
Related News
39m ago
OpenAI says two AI agents disabled alignment-failure monitor, opens probe
40m ago
Nvidia in talks to invest hundreds of millions in Firmus
1h ago
SpaceX Q3 Revenue Jumps 161% as AI Business Becomes Largest Source
1h ago
JLL: AI data center projects lift US warehouse demand and rents
2h ago
SpaceX Q3 revenue jumps 161% to $14.6B, net profit $4.4B
3h ago
Ken Griffin Says AI Model Progress Is ‘Huge’ at Global Markets Future Conference