Simon Willison
Two new speech-to-text models (similar to Whisper) from Mistral today - one of them is API-only, the other is a 8.9GB Apache-2.0 licensed open weights model for "realtime" transcription. They're both very good! https://simonwillison.net/2026/Feb/4/voxtral-2/
Voxtral transcribes at the speed of sound

Mistral just released Voxtral Transcribe 2 - a family of two new models, one open weights, for transcribing audio to text. This is the latest in their Whisper-like model family, …

Simon Willison’s Weblog
Feb 4 at 10:43pmWeb
Show threadMat]3Feb 4
@simon Whisper is very good, but when the audio is noisy and not very clear, it starts hallucinating. I wonder how Voxtral fares in this case. The first Voxtral was just a bit worse than Whisper.
Show threadAndreas WagnerFeb 4

@mathis @simon Whisper seems to have significant hallucination problems with speakers with speech disabilities - even worse than with accents. In a stroke of genius the researchers who investigated this labelled their study "Careless Whisper". I wonder how Voxtral would fare in such situations.

https://doi.org/10.1145/3630106.3658996

Show threadViraptorFeb 5
@simon
It's the first one you can run yourself that does diarization, isn't it? I've seen hacks to implement it that were painful to use before, but nothing truly integrated.
Show threadTim CavellFeb 5
@simon @jsnell maybe something to look into for Apple earnings calls
Show threadSimon DückertFeb 5
@simon Did you compare it to #parkeet v3 in terms of speed and acuracy?