voxtral-mini-3b-2507 (Vokra GGUF)

Speech recognition model, converted to the Vokra GGUF format for Vokra, a zero-dependency speech-AI inference runtime.

This is a conversion, not a new model. The weights are the upstream ones; Vokra re-packages them so its runtime can memory-map them directly. Credit for the model belongs upstream — see Source below.

Files

File	Size	SHA-256
voxtral-mini-3b-2507.gguf	8923.0 MB	eb6c3d317ccc55fa5e4708ad9f80e7d2a145257f9cd4762156fc8ddff45fb233

Usage

# Download (any HTTP client works — the file is a plain GGUF)
curl -L -o voxtral-mini-3b-2507.gguf \
  https://huggingface.co/vokra/voxtral-mini-3b-2507/resolve/main/voxtral-mini-3b-2507.gguf

vokra-cli run --model voxtral-mini-3b-2507.gguf --input speech.wav

# Beam search instead of greedy
vokra-cli run --model voxtral-mini-3b-2507.gguf --input speech.wav --beam-size 5

Prints the transcript. Input must be mono 16 kHz WAV.

Provenance

Field	Value
Architecture	voxtral
Tensors	762
Upstream source	mistralai/Voxtral-Mini-3B-2507 (Apache-2.0)
Upstream licence	apache-2.0
Licence class	permissive
Registry model id	voxtral-mini-3b
Vokra GGUF schema	1
Converted by	vokra-core 0.1.0-alpha.0

Every row above is read out of this file's own vokra.* metadata, so the card cannot claim something the artifact does not carry.

Licence

The weights are distributed under apache-2.0, unchanged from upstream. Conversion does not alter the licence, and your obligations run to the upstream author.

Verifying this file

shasum -a 256 voxtral-mini-3b-2507.gguf
# expect: eb6c3d317ccc55fa5e4708ad9f80e7d2a145257f9cd4762156fc8ddff45fb233