Voxtral Mini · 4B Realtime

Open-weight, multilingual speech transcription with an official streaming vLLM path and an offline-file Transformers example.

Verified sourceRepository opened Jan 21, 2026Source checked 9/23/2026Version: 2769294d

LOCALRENTED GPUOPEN WEIGHTS

At a glance

Parameters: 4B
Architecture: Causal audio encoder and language model
License: Apache 2.0
Software: vLLM, Transformers

View the model source

MY HARDWARE

Will it run on your machine?

Save your machine to see a personalized rating and its reasoning.

Add my hardware

Best for

Transcribe a local audio file without a hosted speech API.
Build private live captions or meeting transcription with the vLLM realtime endpoint.

Tradeoffs

The 16 GB minimum is publisher guidance for the vLLM path; other runtimes and long recordings need separate validation.
Publisher-provided Mac and community runtime paths are less established than the documented CUDA path.

Ways to run it

vLLM · Linux

Repository-specific example documented in the publisher model card; confirm dependencies and platform in the source.

Official instructions

Transformers · See official guide

Repository-specific example documented in the publisher model card; confirm dependencies and platform in the source.

from transformers import VoxtralRealtimeForConditionalGeneration, AutoProcessor
from mistral_common.tokens.tokenizers.audio import Audio
from huggingface_hub import hf_hub_download

repo_id = "mistralai/Voxtral-Mini-4B-Realtime-2602"

processor = AutoProcessor.from_pretrained(repo_id)
model = VoxtralRealtimeForConditionalGeneration.from_pretrained(repo_id, device_map="auto")

repo_id = "patrickvonplaten/audio_samples"
audio_file = hf_hub_download(repo_id=repo_id, filename="bcn_weather.mp3", repo_type="dataset")

audio = Audio.from_file(audio_file, strict=False)
audio.resample(processor.feature_extractor.sampling_rate)

inputs = processor(audio.audio_array, return_tensors="pt")
inputs = inputs.to(model.device, dtype=model.dtype)

outputs = model.generate(**inputs)
decoded_outputs = processor.batch_decode(outputs, skip_special_tokens=True)

print(decoded_outputs[0])

Official instructions

Explore its uses

TRANSCRIPTIONSPEECH RECOGNITIONMULTILINGUALREALTIME

Get it running

Transcribe an audio file privately

Use the publisher’s Transformers path for Voxtral Mini Realtime to transcribe a local recording, check the words against the audio, and keep the file off hosted APIs.

View workflow