Mistral AI released Voxtral 3, an updated multilingual speech-to-text and speech recognition model. The new version features improved accuracy, faster inference speeds, and support for more languages compared to its predecessor.
Background
Mistral AI has been advancing multilingual AI capabilities with models supporting over 100 languages. Voxtral 3 builds on their previous speech-to-text offerings in the competitive open-weight model space.
- Source
- Hacker News (RSS)
- Published
- Oct 21, 2026 at 12:02 AM
- Score
- 8.0 / 10