Mistral AI · Speech to text · 4.43B parameters · Apache-2.0
Voxtral Mini 4B Realtime Arabic
Needs 6 GB at FP8 / INT8. The smallest real configuration today is 1× RTX 3060 at Vast.ai, $0.04 per hour, about $26 a month at 730 hours.
Parameters4.43Bdense
CategorySpeech to textsized on weights + headroom
Checkpoint9 GBBF16, as published
LicenseApache-2.0check before commercial use
ReleasedOct 8, 2026documented date
Smallest configuration · FP8 / INT8
1× RTX 306012 GB VRAM · 44% used
- Per hour
- $0.04
- Per month
- $26
- Cheapest at
Vast.ai
What is documented about Voxtral Mini 4B Realtime Arabic
news checked 2026-10-27Fine-tune of Voxtral-Mini-4B-Realtime-2602 for Arabic dialects, MSA and code-switching (benchmarks co-developed with Morocco's MTNRA). Repo stores the weights twice (consolidated.safetensors and model.safetensors, 8.86 GB each). Mistral's Voxtral 3 family (2026-10-20) is reported as public preview without public weights. Date = HF repo creation.
- https://huggingface.co/mistralai/Voxtral-Mini-4B-Realtime-Arabic
- https://huggingface.co/api/models/mistralai/Voxtral-Mini-4B-Realtime-Arabic?blobs=true
- https://huggingface.co/mistralai/Voxtral-Mini-4B-Realtime-2602
Memory · 3 precisionsHow much VRAM it needs
Weights plus 20% runtime overhead. For speech to text models, activations also depend on resolution, clip length and batch size; the per-GPU estimates below give the practical configuration.
| Precision | Weights | Weights + headroom |
|---|---|---|
| BF16published checkpoint | 9 GB | 11 GB |
| FP8 / INT8estimated | 4 GB | 6 GB |
| INT4 / MXFP4estimated | 3 GB | 4 GB |
Cross-checked on Hugging Face Oct 28, 2026: 4.43B parameters in the safetensors headers, 8.9 GB of weight files. Method: weights as published at the native precision, verified on Hugging Face 2026-10-28; other precisions estimated at 1 (8-bit) and 0.6 (4-bit) bytes per parameter; × 1.2 runtime headroom; KV cache not added (architecture not readable).
Smallest configuration per GPU · 109 GPUs · cheapest eligible on-demand offer per GPU count, every provider on fileWhat it costs to run, GPU by GPU
The fewest GPUs of one type whose total memory fits the model at short context, priced from the cheapest eligible on-demand offer for exactly that count (45-day window, newest reading per provider), with the next providers beside it. Monthly = 730 hours.
| GPU | Smallest config | VRAM used | $ / hour | $ / month | Cheapest at |
|---|---|---|---|---|---|
| RTX 3060Best fit | 1× 12 GB12 GB total | 44% | $0.04 | $26 | Vast.aiTrust 78 · read Nov 2 · receipt · also Clore.ai $0.05, gcube $0.05 |
| RTX A2000 | 1× 12 GB12 GB total | 44% | $0.04 | $26 | Vast.aiTrust 78 · read Nov 2 · receipt · also gcube $0.09, Runpod $0.12 |
| GTX 1660 Ti | 1× 6 GB6 GB total | 89% | $0.04 | $26 | Vast.aiTrust 78 · read Nov 2 · receipt · also Theta EdgeCloud $0.12 |
| Tesla P4 | 1× 8 GB8 GB total | 67% | $0.04 | $29 | Vast.aiTrust 78 · read Nov 2 · receipt · also Akash Network $0.06, gcube $0.07 |
| T4 | 1× 16 GB16 GB total | 33% | $0.04 | $29 | Vast.aiTrust 78 · read Nov 2 · receipt · also iwinv $0.20, Theta EdgeCloud $0.20 |
| GTX 1060 | 1× 6 GB6 GB total | 89% | $0.04 | $30 | Vast.aiTrust 78 · read Nov 2 · receipt |
| GTX 1660 | 1× 6 GB6 GB total | 89% | $0.04 | $30 | Vast.aiTrust 78 · read Nov 2 · receipt · also Edgevana $0.78 |
| Tesla P40 | 1× 24 GB24 GB total | 22% | $0.04 | $30 |  Clore.aiTrust 49 · read Nov 2 · receipt · also gcube $0.07, Vast.ai $0.11 |
| GTX Titan X | 1× 12 GB12 GB total | 44% | $0.04 | $31 | Vast.aiTrust 78 · read Nov 2 · receipt |
| Titan Xp | 1× 12 GB12 GB total | 44% | $0.04 | $31 | Vast.aiTrust 78 · read Nov 2 · receipt |
| GTX 1080 | 1× 8 GB8 GB total | 67% | $0.04 | $32 | Vast.aiTrust 78 · read Nov 2 · receipt · also Edgevana $0.06 |
| Tesla P100 | 1× 16 GB16 GB total | 33% | $0.04 | $33 | Vast.aiTrust 78 · read Nov 2 · receipt · also gcube $0.07 |
| RTX 2060 | 1× 6 GB6 GB total | 89% | $0.04 | $33 | |
| GTX 1070 | 1× 8 GB8 GB total | 67% | $0.05 | $36 | Vast.aiTrust 78 · read Nov 2 · receipt · also Clore.ai $0.08 |
| GTX 1660 Super | 1× 6 GB6 GB total | 89% | $0.05 | $38 |  Theta EdgeCloudnot scored · read Nov 2 · receipt |
| Quadro P4000 | 1× 8 GB8 GB total | 67% | $0.06 | $40 | Vast.aiTrust 78 · read Nov 2 · receipt · also Edgevana $0.08 |
| RTX 2070 | 1× 8 GB8 GB total | 67% | $0.06 | $41 | Vast.aiTrust 78 · read Nov 2 · receipt · also gcube $0.06, Clore.ai $0.21 |
| RTX 3060 Ti | 1× 8 GB8 GB total | 67% | $0.06 | $41 | Vast.aiTrust 78 · read Nov 2 · receipt · also Clore.ai $0.06, gcube $0.06 |
| RTX 2060 Super | 1× 8 GB8 GB total | 67% | $0.06 | $41 | |
| Quadro P2000 | 2× 5 GB10 GB total | 53% | $0.06 | $41 | Vast.aiTrust 78 · read Nov 2 · receipt |
| GTX 1070 Ti | 1× 8 GB8 GB total | 67% | $0.06 | $44 | Vast.aiTrust 78 · read Nov 2 · receipt · also Clore.ai $0.10 |
| RTX 2070 Super | 1× 8 GB8 GB total | 67% | $0.06 | $44 | |
| RTX 3070 | 1× 8 GB8 GB total | 67% | $0.06 | $44 | |
| RTX 4060 | 1× 8 GB8 GB total | 67% | $0.06 | $45 | Vast.aiTrust 78 · read Nov 2 · receipt · also Clore.ai $0.08, Theta EdgeCloud $0.09 |
| The 24 cheapest of 106 GPU types that fit. Every GPU, any precision and context, in the finder | |||||
Voxtral Mini 4B Realtime Arabic · related