available
Weights
Official weights are downloadable.
Inspect this claim’s source ↗Mistral AI / FAMILY RECORD
Voxtral Realtime: license, size and what is open.
A compact streaming speech-recognition model designed for low-latency transcription through a realtime API.
Sizes and download
4B
- Voxtral-Mini-4B-Realtime-2602
- Total: 4B. Active: 4B. Source ↗
About 8.86 GB of selected documented weights · File listing ↗
Selected artifact: mistralai/Voxtral-Mini-4B-Realtime-2602. Format: Safetensors, Mistral consolidated representation. File evidence checked 2026-10-11. Selected artifact license: apache-2.0 ↗
Single consolidated.safetensors. Excludes alternative model.safetensors (8,859,446,848 bytes); alternate representations are not complementary shards. Selected weight files only; supporting files and runtime memory are separate. Not a smallest-download census or proof of hardware compatibility.
Can my computer run it?
Compatibility, runtime memory and speed are untested by Bright.
Choose your computer and inspect documented candidates →Nine things to check
Source check: . Downloads, when known, carry their own check date.
Checked release: Voxtral-Mini-4B-Realtime-2602
Sources checked 2026-09-19
- Voxtral Realtime release
- · Primary source
The profile describes this checkpoint, not every release carrying the family name. Developer documentation is not independent reproduction.
THE MODEL / THE TERMS
Know what you are picking up.
- Parameters
- 4B
- Architecture
- Streaming audio encoder-decoder model integrated with Mistral inference runtimes.
- Modalities
- audio · text
- Context
- Streaming sessions; a comparable language-token window is not asserted here.
- Weights license
- Apache-2.0
- Commercial use
- Permitted by Apache-2.0.
- Memory
- The 4B size supports comparatively modest deployment, but streaming overhead and audio session length affect requirements.
- Known quantizations
- Runtime-supported formats should be verified against the official card
Specifications are developer-reported. Primary release documentation ↗ · All sources and terms
NINE DIMENSIONS / NO OPENNESS SCORE
What is actually available?
Each component has its own evidence. “Not established” means this source review did not verify availability; it is not a claim that nothing exists.
available
Architecture
Configuration and model card are public.
Inspect this claim’s source ↗available
Inference code
The release documents vLLM’s Realtime API integration.
Inspect this claim’s source ↗not established
Training code
Complete training code was not confirmed.
Inspect this claim’s source ↗not established
Training recipe
A full reproducible training recipe was not confirmed.
Inspect this claim’s source ↗partial
Data information
Capabilities and languages are described without a complete corpus inventory.
Inspect this claim’s source ↗not established
Training data
No sufficiently specific public artifact was confirmed in this review.
Inspect this claim’s source ↗partial
Evaluation
Vendor speech evaluations are published in the model materials.
Inspect this claim’s source ↗available
Commercial use
The official checkpoint is Apache-2.0 licensed.
Inspect this claim’s source ↗WHY THIS RELEASE MATTERS
What to explore.
- Streaming transcription
- Low-latency audio processing
- Compact size
WHAT THE RECORD DOES NOT ESTABLISH
Keep the limits.
- Speech-focused rather than a general multimodal assistant
- Training data is not fully disclosed
vLLM integration corroborates production serving support; transcription quality claims are not independently reproduced here.
FOLLOW THE RECORD
Sources, not superlatives.
- Voxtral Transcribe 2 and Realtime release ↗ (institution)
- Voxtral Mini 4B Realtime model card ↗ (registry)
- Mistral model changelog ↗ (institution)
Source check: 2026-09-19. AI-assisted editorial review, with no named external expert or independent hardware validation claimed. Licenses and model cards can change. Review the linked terms for the specific artifact and intended use.
Understand open source vs open weights ↗