GR

Grok 1

GrokGGUF by Ozias1111Updated 2026-02-22
Download on Hugging Face
Parameters316B
Context8K
Layers64
Architecturegrok
Smallest quant65.3 GB
Quick answer

Grok 1 needs about 139.8 GB of memory at IQ3_S (a 137.1 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 25 quants, from 65.3 GB (IQ1_S) to 335.4 GB (Q8_0).

Grok 1 is a 316B-parameter Grok model in GGUF format, quantized by Ozias1111. Its trained context length is 8K tokens. The repository has 25 quant files, from 65.3 GB (IQ1_S) to 335.4 GB (Q8_0).

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
IQ1_S65.3 GB (9 parts) 67.9 GB
Too big Get file →
IQ1_M72.3 GB (9 parts) 74.9 GB
Too big Get file →
IQ2_XXS83.9 GB (9 parts) 86.6 GB
Too big Get file →
IQ2_XS93.3 GB (9 parts) 95.9 GB
Too big Get file →
IQ2_S94.6 GB (9 parts) 97.2 GB
Too big Get file →
IQ2_M103.9 GB (9 parts) 106.6 GB
Too big Get file →
Q2_K_S107.6 GB (9 parts) 110.3 GB
Too big Get file →
Q2_K115.9 GB (9 parts) 118.6 GB
Too big Get file →
Q2_K116.2 GB (9 parts) 118.8 GB
Too big Get file →
IQ3_XXS122.2 GB (9 parts) 124.9 GB
Too big Get file →
IQ3_XS129.6 GB (9 parts) 132.3 GB
Too big Get file →
IQ3_XS129.8 GB (9 parts) 132.4 GB
Too big Get file →
IQ3_S137.1 GB (9 parts) 139.8 GB
Too big Get file →
Q3_K_S137.1 GB (9 parts) 139.8 GB
Too big Get file →
IQ3_M144.0 GB (9 parts) 146.6 GB
Too big Get file →
Q3_K_M151.6 GB (9 parts) 154.3 GB
Too big Get file →
Q3_K_L163.1 GB (9 parts) 165.7 GB
Too big Get file →
IQ4_XS168.8 GB (9 parts) 171.4 GB
Too big Get file →
IQ4_NL178.5 GB (9 parts) 181.2 GB
Too big Get file →
Q4_K191.8 GB (9 parts) 194.4 GB
Too big Get file →
Q4_K192.2 GB (9 parts) 194.9 GB
Too big Get file →
Q5_K224.3 GB (9 parts) 227.0 GB
Too big Get file →
Q6_K259.2 GB (9 parts) 261.8 GB
Too big Get file →
Q6_K259.8 GB (9 parts) 262.5 GB
Too big Get file →
Q8_0335.4 GB (9 parts) 338.1 GB
Too big Get file →

What GPU do you need for Grok 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 Grok 1 quant should you choose?

IQ3_S (137.1 GB) is the middle option in this repository. A reasonable choice when a 4-bit file will not fit, mostly on larger models. Quality loss is noticeable but often acceptable.

If memory is tight, the smallest file is IQ1_S at 65.3 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

Architecturegrok
Parameters316B (approximate, mixture of experts)
Layers64
Trained context8,192 tokens
KV cache (fp16)256 KB per token

Source

Quantized byOzias1111
FamilyGrok
Licenseapache-2.0
Downloads (HF)83
Last updated2026-02-22

Frequently asked questions about Grok 1

How much VRAM does Grok 1 need?

About 139.8 GB at IQ3_S with an 8K context window. That is the 137.1 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 Grok 1 on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 67.9 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Grok 1 GGUF?

It depends on the quant: IQ1_S is 65.3 GB, IQ1_M is 72.3 GB, IQ2_XXS is 83.9 GB, IQ2_XS is 93.3 GB, IQ2_S is 94.6 GB, IQ2_M is 103.9 GB and 19 more.

What context length does Grok 1 support?

The GGUF header lists a trained context length of 8,192 tokens (8K). The KV cache grows with the context you use: about 2.1 GB at 8K and 2.1 GB at 8K.

Where can I download Grok 1 GGUF?

The files are on Hugging Face in the Ozias1111/Grok-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 Grok 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.

Similar models

Browse all →

More Grok models

All Grok →

Other large models

All large models →