QW

QwQ 56B Ghost

QwenGGUF by mradermacherUpdated 2025-01-20
Download on Hugging Face
Parameters56.2B
Context32K
Layers112
Architectureqwen2
Smallest quant12.1 GB
Quick answer

QwQ 56B Ghost needs about 38.2 GB of memory at Q4_K_M (a 33.9 GB file) with an 8K context window, so it fits on a 48 GB GPU. This repository has 23 quants, from 12.1 GB (IQ1_S) to 46.1 GB (Q6_K). Longer context needs more: at 32K tokens the KV cache alone adds about 15.0 GB.

QwQ 56B Ghost is a 56.2B-parameter Qwen model in GGUF format, quantized by mradermacher. Its trained context length is 32K tokens. The repository has 23 quant files, from 12.1 GB (IQ1_S) to 46.1 GB (Q6_K).

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
IQ1_S12.1 GB 16.4 GB
Too big Get file →
IQ1_M13.3 GB 17.5 GB
Too big Get file →
IQ2_XXS15.2 GB 19.5 GB
Too big Get file →
IQ2_XS16.8 GB 21.1 GB
Too big Get file →
IQ2_S17.5 GB 21.8 GB
Too big Get file →
IQ2_M19.1 GB 23.3 GB
Too big Get file →
Q2_K_S19.4 GB 23.7 GB
Too big Get file →
Q2_K20.9 GB 25.1 GB
Too big Get file →
IQ3_XXS21.8 GB 26.1 GB
Too big Get file →
IQ3_XS23.3 GB 27.5 GB
Too big Get file →
Q3_K_S24.5 GB 28.7 GB
Too big Get file →
IQ3_S24.5 GB 28.8 GB
Too big Get file →
IQ3_M25.2 GB 29.4 GB
Too big Get file →
Q3_K_M27.2 GB 31.4 GB
Too big Get file →
Q3_K_L29.4 GB 33.7 GB
Too big Get file →
IQ4_XS30.2 GB 34.4 GB
Too big Get file →
Q4_031.9 GB 36.2 GB
Too big Get file →
Q4_K_S32.1 GB 36.3 GB
Too big Get file →
Q4_K_M33.9 GB 38.2 GB
Too big Get file →
Q4_135.3 GB 39.5 GB
Too big Get file →
Q5_K_S38.7 GB 43.0 GB
Too big Get file →
Q5_K_M39.8 GB 44.1 GB
Too big Get file →
Q6_K46.1 GB 50.3 GB
Too big Get file →

What GPU do you need for QwQ 56B Ghost?

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: the best quant that fits is Q2_K_S (19.4 GB file, about 23.7 GB in use).
  • 32 GB VRAM: the best quant that fits is Q3_K_M (27.2 GB file, about 31.4 GB in use).
  • 48 GB VRAM: the best quant that fits is Q5_K_M (39.8 GB file, about 44.1 GB in use).

Which QwQ 56B Ghost quant should you choose?

Q4_K_M is the usual starting point for QwQ 56B Ghost: it balances file size and quality well. Here it is a 33.9 GB file.

If memory is tight, the smallest file is IQ1_S at 12.1 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

Architectureqwen2
Parameters56.2B
Layers112
Trained context32,768 tokens
KV cache (fp16)448 KB per token

Source

Quantized bymradermacher
FamilyQwen
Licenseapache-2.0
Downloads (HF)93
Last updated2025-01-20

Frequently asked questions about QwQ 56B Ghost

How much VRAM does QwQ 56B Ghost need?

About 38.2 GB at Q4_K_M with an 8K context window. That is the 33.9 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 QwQ 56B Ghost on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 16.4 GB, so you would need a larger GPU or CPU offloading.

What is the file size of QwQ 56B Ghost GGUF?

It depends on the quant: IQ1_S is 12.1 GB, IQ1_M is 13.3 GB, IQ2_XXS is 15.2 GB, IQ2_XS is 16.8 GB, IQ2_S is 17.5 GB, IQ2_M is 19.1 GB and 17 more.

What context length does QwQ 56B Ghost support?

The GGUF header lists a trained context length of 32,768 tokens (32K). The KV cache grows with the context you use: about 3.8 GB at 8K and 15.0 GB at 32K.

Where can I download QwQ 56B Ghost GGUF?

The files are on Hugging Face in the mradermacher/QwQ-56B-Ghost-i1-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 QwQ 56B Ghost 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 from mradermacher

All by mradermacher →

More Qwen models

All Qwen →

Other large models

All large models →