Higgs Audio v3 STT and VibeVoice-ASR
Microsoft's free-to-download speech-to-text model that can transcribe up to an hour of audio, labelling who spoke and when, for developers building transcription tools.
trckr found this on Nov 15, 250 days after it came out, so it sent no alert. Its release date comes from Hugging Face.
Key facts
- Released
- Mar 10, 2026
- What it does
- Makes sound: Creates speech, music or other audio.
Can I use it?
In an app
No app mentioned so far.
For developers
No developer access mentioned so far.
Download
Free to download and run on your own computer: it has Open weights: The model's files are public, so anyone can download it, run it on their own computer and build on it, usually for free. trckr labels these "Open"..
Compare Higgs Audio v3 STT and VibeVoice-ASR with
It came out 194 days after MAI-Voice-1, the lab's previous model.
The details
VibeVoice-ASR is an 8.3 billion parameters: Runs on a good laptop or desktop computer. Parameters are the numbers a model learns: a rough measure of its size. Speech-to-text: Turns spoken audio into written text. Also called transcription or ASR. model from Microsoft that transcribes up to 60 minutes of continuous audio in one pass, producing structured who-said-what-and-when output. It is released as free-to-download weights under the MIT license, usable through Hugging Face: The website where labs and researchers publish models anyone can download. Think of it as an app store for AI models. Transformers.
Model ID: The exact name developers type to use a model, like gpt-5-mini. New IDs often show up in a lab's systems before the official announcement, which is how trckr spots many releases early.
bosonai/higgs-audio-v3-sttmicrosoft/VibeVoice-ASR-HFTimeline
/GitHub ggml-org/llama.cppFirst signal
mtmd: support higgs v3 STT and vibevoice-ASR (ggml-org/llama.cpp #32329)
Get alerts like this the moment they happen.
From Microsoft and every other major lab.