MI

Mixtral 8x22B v0.1

MistralGGUF by LiteLLMsUpdated 2024-04-18
Download on Hugging Face
Parameters141B
Context64K
Layers56
Architecturellama
Smallest quant52.1 GB
Quick answer

Mixtral 8x22B v0.1 needs about 88.0 GB of memory at Q4_K_M (a 85.6 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 18 quants, from 52.1 GB (Q2_K) to 115.5 GB (Q6_K). Longer context needs more: at 32K tokens the KV cache alone adds about 7.5 GB.

Mixtral 8x22B v0.1 – a 141‑billion‑parameter mixture‑of‑experts model in GGUF format

Mixtral 8x22B v0.1 is a 141‑billion‑parameter mixture‑of‑experts model based on the Mistral family. It supports a 65536‑token context window and is distributed in multiple GGUF quantizations. The model is released under the apache‑2.0 license.

Good at

  • Large context handling
  • High‑capacity reasoning
  • Efficient inference with quantization

Limits

  • No vision support
  • Requires significant VRAM for higher quantizations
  • No active parameters listed

Best for

  • Long‑form content generation
  • Complex question answering
  • Research prototyping

Who it’s for. Developers and researchers needing a large‑scale language model that can be run locally with llama.cpp or compatible tools. The model is suitable for those who can manage the memory requirements of a 141‑billion‑parameter architecture.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q2_K52.1 GB (9 parts) 54.5 GB
Too big Get file →
IQ3_XS58.2 GB (9 parts) 60.6 GB
Too big Get file →
IQ3_S61.5 GB (9 parts) 63.9 GB
Too big Get file →
Q3_K_S61.5 GB (9 parts) 63.9 GB
Too big Get file →
IQ3_M64.5 GB (9 parts) 66.9 GB
Too big Get file →
Q3_K_M67.8 GB (9 parts) 70.2 GB
Too big Get file →
IQ4_XS76.4 GB (9 parts) 78.7 GB
Too big Get file →
Q3_K_L76.6 GB (10 parts) 79.0 GB
Too big Get file →
Q4_079.5 GB (9 parts) 81.9 GB
Too big Get file →
IQ4_NL80.5 GB (9 parts) 82.9 GB
Too big Get file →
Q4_K_S80.5 GB (9 parts) 82.9 GB
Too big Get file →
Q4_K_M85.6 GB (9 parts) 88.0 GB
Too big Get file →
Q4_188.2 GB (9 parts) 90.6 GB
Too big Get file →
Q5_097.0 GB (9 parts) 99.4 GB
Too big Get file →
Q5_K_S97.0 GB (9 parts) 99.4 GB
Too big Get file →
Q5_K_M100.0 GB (9 parts) 102.3 GB
Too big Get file →
Q5_1105.7 GB (9 parts) 108.1 GB
Too big Get file →
Q6_K115.5 GB (9 parts) 117.9 GB
Too big Get file →

For 8 GB VRAM, use Q2_K (52.1 GB file, ~54.5 GB memory at 8 k context). For 16 GB VRAM, Q3_K_S or IQ3_S (61.5 GB file, ~63.9 GB memory). For 24 GB VRAM, Q4_K_M (85.59 GB file, ~88 GB memory). Memory estimates are approximate.

What GPU do you need for Mixtral 8x22B v0.1?

These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.

  • 4 GB VRAM: no quant fits at an 8K context.
  • 6 GB VRAM: no quant fits at an 8K context.
  • 8 GB VRAM: no quant fits at an 8K context.
  • 12 GB VRAM: no quant fits at an 8K context.
  • 16 GB VRAM: no quant fits at an 8K context.
  • 24 GB VRAM: no quant fits at an 8K context.
  • 32 GB VRAM: no quant fits at an 8K context.
  • 48 GB VRAM: no quant fits at an 8K context.

Which Mixtral 8x22B v0.1 quant should you choose?

Q4_K_M is the usual starting point for Mixtral 8x22B v0.1: it balances file size and quality well. Here it is a 85.6 GB file.

If memory is tight, the smallest file is Q2_K at 52.1 GB, with the quality trade-offs that come with fewer bits. Use it only when memory is very tight. Expect a clear drop in quality, which is usually worse on small models and easier to live with on large ones.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturellama
Parameters141B (approximate, mixture of experts)
Layers56
Trained context65,536 tokens
KV cache (fp16)224 KB per token

Source

Quantized byLiteLLMs
FamilyMistral
Licenseapache-2.0
Downloads (HF)2,088
Last updated2024-04-18

Frequently asked questions about Mixtral 8x22B v0.1

How much VRAM does Mixtral 8x22B v0.1 need?

About 88.0 GB at Q4_K_M with an 8K context window. That is the 85.6 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.

Can I run Mixtral 8x22B v0.1 on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 54.5 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Mixtral 8x22B v0.1 GGUF?

It depends on the quant: Q2_K is 52.1 GB, IQ3_XS is 58.2 GB, IQ3_S is 61.5 GB, Q3_K_S is 61.5 GB, IQ3_M is 64.5 GB, Q3_K_M is 67.8 GB and 12 more.

What context length does Mixtral 8x22B v0.1 support?

The GGUF header lists a trained context length of 65,536 tokens (64K). The KV cache grows with the context you use: about 1.9 GB at 8K and 7.5 GB at 32K.

Where can I download Mixtral 8x22B v0.1 GGUF?

The files are on Hugging Face in the LiteLLMs/Mixtral-8x22B-v0.1-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

What license does Mixtral 8x22B v0.1 use?

Hugging Face lists the license as apache-2.0. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.

Can I run Mixtral 8x22B v0.1 on 8 GB VRAM?

Only the lowest‑bit quantizations such as Q2_K are feasible, but memory usage remains high and may exceed 8 GB in practice.

Does the model support vision tasks?

No, the model has no vision capabilities.

What license applies to this model?

The model is released under the apache‑2.0 license.

Similar models

Browse all →

More from LiteLLMs

All by LiteLLMs →

More Mistral models

All Mistral →

Other large models

All large models →