mistralai / automatic-speech-recognition updated 4 months ago

Voxtral-Mini-4B-Realtime-2602

Name: Voxtral-Mini-4B-Realtime-2602
Author: mistralai

Voxtral Mini 4B Realtime 2602 is a multilingual, realtime speech-transcription model and among the first open-source solutions to achieve accuracy comparable to offline systems with a delay of <500ms. It supports 13 languages and outperforms existing open-source baselines across a range of tasks, making it ideal for applications like voice assistants and live subtitling.

Params

4.4 B

Context

—

Downloads 30d

2.3 M

Likes

936

Commercial use: allowed apache-2.0 Not gated SAFETENSORS 13 languages View on Hugging Face ↗

Download history

daily snapshots · 10 days

2.3 M2.0 M

Jul 28Jul 31Aug 3Aug 6

Can you run it?

Estimated VRAM at 8K context unless noted. Pick your hardware to see the verdict per quantization.

File	Quant	Size	Est. VRAM	Verdict on RTX 4090 · 24 GB
model.safetensors	bf16	17.7 GB	20.7 GB	⚠️ Tight — reduce context

Estimate: file size × 1.1 + KV cache at 8K + 0.5 GB overhead. Not a benchmark — how we calculate this.

Specifications

Architecture: VoxtralRealtimeForConditionalGeneration
Parameters: 4.4 B
Tensor type: BF16
Licence: apache-2.0
First seen on the Hub: 2026-01-21
Training datasets: undisclosed
Added to our catalog: 2026-07-28

Family

Base model and the most-downloaded derivatives in the catalog.

Derivative Voxtral-Mini-4B-Realtime-2602-gguf handy-computer 430 K

Compare with

speaker-diarization-3.1 8.9 M whisperkit-coreml 8.3 M speaker-diarization-community-1 5.3 M wav2vec2-large-xlsr-53-portuguese 4.6 M wav2vec2-large-xlsr-53-russian 4.0 M voice-activity-detection 2.7 M

Appears in

Best open coding models Commercial-use text models Models that fit in 12 GB