How much VRAM does Mixtral 8x7B Instruct need?

About 30.1 GB at Q4_K_M with 4K of context and batch 1, rising to roughly 34.0 GB at 32K. That figure is weights plus KV cache plus a 10% activation buffer, calculated from the model's architecture rather than measured on a card.

Will Mixtral 8x7B Instruct run on a 40GB GPU?

Yes — at Q4_K_M and 4K context it needs about 30.1 GB, which leaves 9.9 GB spare on a A100 40G. That is the smallest card in this index that clears it comfortably; 19 of 63 do.

Which quantization of Mixtral 8x7B Instruct should I use?

Q4_K_M has the lowest published quality loss (2.8%), and Q4_K_M is the level most people run. All 2 levels in the index are Q4_K_M, AWQ INT4.

Mixtral 8x7B Instruct

47B MoE

Mistral AI

⚠ Superseded · Prefer Qwen3 30B-A3B Instruct

This index would suggest Qwen3 30B-A3B Instruct instead today — longer context (32K → 40K tokens). The numbers below are still accurate; they are just for a model you probably should not start with. Compare the two →

Classic MoE model. ~13B active params per token; needs 32GB+ VRAM for Q4.

Pro GPU

33K

Max Context

Quant Variants

GGUF Q4_K_M

Best Quality

97.2%

Accuracy Retained

Calculate VRAM Hugging Face Compare

Quantization Variants

Per-quant VRAM, quality loss, and inference speed on RTX 4090

Measured = site benchmarks · Estimated = formula · Community = public reports

Format	Level	BPW	VRAM	PPL Loss	Speed	Source	Actions
GGUF	Q4_K_M	4.85	28.5 GB	2.8%	48 tok/s	Estimated	Calc HF
AWQ	INT4	4	25.2 GB	3.8%	62 tok/s	Estimated	Calc HF

Similar models

Compare with Mistral Large

675B MoE

Mistral Large 3 675B Instruct

Mistral AI

Pro GPU

312.0 GBmin VRAM97.9%accuracy

Mistral 3 flagship MoE (41B active / 675B total) with vision encoder. FP8 on 8×H200; GGUF quant for research clusters only.

24B

Mistral Small 24B Instruct

Mistral AI

Consumer GPUPro GPU

13.5 GBmin VRAM97.8%accuracy

Mistral's efficient 24B. Strong multilingual; fits on 24GB with Q4.

12B

Mistral Nemo 12B Instruct

Mistral AI

Consumer GPUMac / Apple Silicon

7.8 GBmin VRAM99.1%accuracy

Mistral + NVIDIA collaboration. 128K context, excellent multilingual support.

Mistral 7B Instruct v0.3

Mistral AI

Consumer GPUMac / Apple Silicon

4.6 GBmin VRAM99.1%accuracy

Classic Mistral 7B v0.3. Still a reliable baseline for local chat APIs.

Running Mixtral 8x7B Instruct locally

At Q4_K_M and 4K of context, Mixtral 8x7B Instruct needs about 30.1 GB — 26.9 GB of weights, 0.50 GB of KV cache and a 2.7 GB activation buffer. The smallest card in this index that clears that comfortably is the A100 40G at 40 GB, and 19 of the 63 cards here do. These are calculated figures, not measurements: the estimate stops counting a card as comfortable at 88% of its VRAM, which is roughly the room a desktop session needs.

What longer context costs

Going from 4K to 32K adds about 3.50 GB, taking the total to 34.0 GB. Weights do not move with context — only the KV cache does, and it grows linearly, so this is the number to watch when planning for long documents. The model's native window is 32K; holding all of it at Q4_K_M would need about 34 GB.

Which build to download

This index tracks 2 formats for it — GGUF, AWQ — across 2 levels. Q4_K_M carries the lowest published perplexity loss at 2.8%. The fastest level measured here is AWQ INT4 at 62 tok/s on an RTX 4090, batch 1. GGUF runs on llama.cpp and Ollama across NVIDIA, AMD and Apple silicon; AWQ and GPTQ target vLLM on CUDA and ROCm; EXL2 is ExLlamaV2 and CUDA only.

Common questions

How much VRAM does Mixtral 8x7B Instruct need?: About 30.1 GB at Q4_K_M with 4K of context and batch 1, rising to roughly 34.0 GB at 32K. That figure is weights plus KV cache plus a 10% activation buffer, calculated from the model's architecture rather than measured on a card.
Will Mixtral 8x7B Instruct run on a 40GB GPU?: Yes — at Q4_K_M and 4K context it needs about 30.1 GB, which leaves 9.9 GB spare on a A100 40G. That is the smallest card in this index that clears it comfortably; 19 of 63 do.
Which quantization of Mixtral 8x7B Instruct should I use?: Q4_K_M has the lowest published quality loss (2.8%), and Q4_K_M is the level most people run. All 2 levels in the index are Q4_K_M, AWQ INT4.

Where this model fits

Sized at Q4_K_M with a 4K context window, smallest card first. Comfortable means the estimate uses at most 88% of the memory.

Runs comfortably on: A100 40G40GB · +9.9 L40S 48G48GB · +17.9 A40 48G48GB · +17.9 Mac M4 Pro 48G48GB · +5.9 Mac M3 Max 48G48GB · +5.9
Tight but possible: RTX 509032GB Instinct MI100 32G32GB
Just misses: Mac M4 Max 36G+3.1GB over Mac M3 Pro 36G+3.1GB over
Ships in: GGUF AWQ
Compared here: GGUF vs AWQ
Try it yourself: Size it in the calculator Generate the run command

How to actually run this

Deployment guides for this model and this class of hardware.

This page's figures change when the model or the runtime does.Last updated 2026-09-21 RSS → /feed.xml