Mistral AI8B~6 GB VRAM at Q4_K_M

Ministral 3 8B — VRAM, speed & local setup

Written by Jakub Rusinowski · Last updated

The balanced middle of the Ministral 3 range at 5.6 GB Q4_K_M — comfortable on an 8 GB card and the natural replacement for a Llama 3.1 8B install. Vision-capable, 128K context, Apache 2.0.

Ministral 3 8B needs about 6 GB of VRAM at Q4_K_M — quantized weights plus framework overhead, before any KV cache. On Apple Silicon that figure comes out of unified memory.

Logic78
Creative76
Coding75

VRAM and speed by quantization

Quoted against NVIDIA RTX 4090 (24 GB). Weights plus overhead only: this model's architecture is unpublished, so no KV cache is added.

QuantVRAMSpeed (est.)Fit
Q2_K
2.63 bpw
3.4 GB~154 tok/sFits
Q3_K_M
3.41 bpw
4.2 GB~133 tok/sFits
Q4_K_M
4.83 bpw
5.6 GB~107 tok/sFits
Q5_K_M
5.67 bpw
6.5 GB~95 tok/sFits
Q6_K
6.56 bpw
7.4 GB~86 tok/sFits
Q8_0
8.50 bpw
9.3 GB~70 tok/sFits
F16
16.00 bpw
16.8 GB~41 tok/sFits

Black marker = usable memory on the NVIDIA RTX 4090 (24 GB). Estimates from the memory-bandwidth roofline on the methodology page. Ministral 3 8B VRAM calculator →

Get Ministral 3 8B running

The cheapest catalogued GPU that runs Ministral 3 8B is the Intel Arc B570 (10 GB).

Buy this hardware Intel Arc B570 10GB — 10 GB VRAM · 150 W board powerDeploy in the cloud now RTX 4090 on RunPod — from $0.34/hr · rate checked 2026-07

or compare on Vast.ai from $0.35/hr (typical low · varies)

As an Amazon Associate we earn from qualifying purchases. Cloud GPU links are referral links — we may earn a commission at no extra cost to you.

Affiliate disclosure: Some links on this page are affiliate links — if you buy through them, LLM Configurator may earn a commission at no extra cost to you. As an Amazon Associate, LLM Configurator earns from qualifying purchases.
Intel Arc B570 10GB
10 GB VRAM · 150 W board power
2026 prices are volatile — check the current listing.

How to run Ministral 3 8B

Install Ollama, then run:

ollama run ministral-3:8b
Weights on Hugging Face: mistralai/Ministral-3-8B-Instruct ↗

Specifications

Verified — Checked against the primary source — the model card or the vendor spec page — and corroborated by a second independent source.

Parameters
8 Billion
Context window
131,072
Architecture
Dense Transformer (vision)
Provider
Mistral AI
Licence
Apache 2.0
Practical deployment
Laptop
Specified at
Q4_K_M
System RAM
16 GB
Record updated
2026-09-06
LicenceApache-2.0Commercial use permitted

Commercial use permitted. No usage restrictions beyond attribution.

Quality and use cases

Scores as published by the model’s authors or an independent evaluator — quality, not throughput, and not measured by us.

Best forgeneral purposeedge devicesmultimodalchat

Can I run Ministral 3 8B on my GPU?

Other Ministral 3 sizes

Ministral 3 8B — frequently asked questions

How much VRAM does Ministral 3 8B need?

About 6 GB at Q4_K_M — quantized weights plus framework overhead, before any KV cache. The cache grows with context length and is added on top; the table above folds it in. Apple Silicon counts unified memory toward the same figure.

Does Ministral 3 8B run on an RTX 4090 (24 GB)?

Yes. Ministral 3 8B needs about 6 GB at Q4_K_M, inside a 24 GB card, at an estimated 107 tokens/sec.

How do I run Ministral 3 8B locally?

Install Ollama and run `ollama run ministral-3:8b`. That pulls the weights and starts a local OpenAI-compatible endpoint; after the download nothing leaves the machine.

What other sizes does Ministral 3 come in?

Ministral 3 3B (3 GB), Ministral 3 8B (6 GB), Ministral 3 14B (9 GB). Every size shares the family's training and licence; the larger ones score higher and need proportionally more memory.