WA

Wan21 FastMix

WanGGUF by ZuntanUpdated 2025-07-26
Download on Hugging Face
Parameters16.4B
Context—
Layers—
Architecturewan
Smallest quant8.1 GB
Quick answer

Wan21 FastMix needs about 12.7 GB of memory at Q4_K_M (a 10.8 GB file) with an 8K context window, so it fits on a 16 GB GPU. This repository has 31 quants, from 8.1 GB (Q3_K_M) to 34.7 GB (F16).

Wan21 FastMix 16.4B parameter model for image‑to‑video generation

Wan21 FastMix is a 16.4B parameter model designed for image‑to‑video synthesis. It supports 4‑step generation with lcm or flowmatch_pusa schedulers and includes LoRA options for 480p and 720p output. The model is available in multiple GGUF quantizations.

Good at

  • Image‑to‑video synthesis
  • Low‑step generation
  • LoRA fine‑tuning

Limits

  • No vision support beyond image‑to‑video
  • No explicit license information
  • Large VRAM requirement for higher quants

Best for

  • Creating short video clips from images
  • Rapid prototyping of video generation pipelines
  • Fine‑tuning with LoRA for specific resolutions

Who it’s for. Developers and researchers working on image‑to‑video pipelines who need a compact 16.4B model with flexible quantization options. The model is suitable for those who can manage the VRAM demands of higher‑precision variants.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q3_K_M8.1 GB 9.6 GB
Too big Get file →
Q3_K_M8.1 GB 9.6 GB
Too big Get file →
Q3_K_M8.1 GB 9.6 GB
Too big Get file →
Q3_K_M8.1 GB 9.6 GB
Too big Get file →
Q3_K_M8.1 GB 9.6 GB
Too big Get file →
Q3_K_M8.6 GB 10.2 GB
Too big Get file →
Q4_K_S9.9 GB 11.6 GB
Too big Get file →
Q4_K_M10.8 GB 12.7 GB
Too big Get file →
Q4_K_M10.8 GB 12.7 GB
Too big Get file →
Q4_K_M10.8 GB 12.7 GB
Too big Get file →
Q4_K_M10.8 GB 12.7 GB
Too big Get file →
Q4_K_M10.8 GB 12.7 GB
Too big Get file →
Q4_K_M11.6 GB 13.5 GB
Too big Get file →
Q6_K13.7 GB 15.9 GB
Too big Get file →
Q6_K13.7 GB 15.9 GB
Too big Get file →
Q6_K13.7 GB 15.9 GB
Too big Get file →
Q6_K13.7 GB 15.9 GB
Too big Get file →
Q6_K13.7 GB 15.9 GB
Too big Get file →
Q6_K14.5 GB 16.8 GB
Too big Get file →
Q8_017.7 GB 20.3 GB
Too big Get file →
Q8_017.7 GB 20.3 GB
Too big Get file →
Q8_017.7 GB 20.3 GB
Too big Get file →
Q8_017.7 GB 20.3 GB
Too big Get file →
Q8_017.7 GB 20.3 GB
Too big Get file →
Q8_018.7 GB 21.4 GB
Too big Get file →
F1632.8 GB 37.2 GB
Too big Get file →
F1632.8 GB 37.2 GB
Too big Get file →
F1632.8 GB 37.2 GB
Too big Get file →
F1632.8 GB 37.2 GB
Too big Get file →
F1632.8 GB 37.2 GB
Too big Get file →
F1634.7 GB 39.3 GB
Too big Get file →

For 8 GB VRAM, use Q3_K_M (8.09 GB file, ~9.6 GB memory). For 12 GB VRAM, Q4_K_M (10.85 GB file, ~12.7 GB memory) is a good balance. If 16 GB VRAM is available, Q6_K (13.74 GB file, ~15.9 GB memory) offers higher precision.

What GPU do you need for Wan21 FastMix?

These figures add the file size, an fp16 KV cache at an 8K context window 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: the best quant that fits is Q4_K_S (9.9 GB file, about 11.6 GB in use).
  • 16 GB VRAM: the best quant that fits is Q6_K (13.7 GB file, about 15.9 GB in use).
  • 24 GB VRAM: the best quant that fits is Q8_0 (18.7 GB file, about 21.4 GB in use).
  • 32 GB VRAM: the best quant that fits is Q8_0 (18.7 GB file, about 21.4 GB in use).
  • 48 GB VRAM: the best quant that fits is F16 (34.7 GB file, about 39.3 GB in use).

Which Wan21 FastMix quant should you choose?

Q4_K_M is the usual starting point for Wan21 FastMix: it balances file size and quality well. Here it is a 10.8 GB file.

The Q8_0 file is 17.7 GB, about 1.6 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.

If memory is tight, the smallest file is Q3_K_M at 8.1 GB, with the quality trade-offs that come with fewer bits. A reasonable choice when a 4-bit file will not fit, mostly on larger models. Quality loss is noticeable but often acceptable.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturewan
Parameters16.4B

Source

Quantized byZuntan
FamilyWan
Downloads (HF)648
Last updated2025-07-26

Frequently asked questions about Wan21 FastMix

How much VRAM does Wan21 FastMix need?

About 12.7 GB at Q4_K_M with an 8K context window. That is the 10.8 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 Wan21 FastMix on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 9.6 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Wan21 FastMix GGUF?

It depends on the quant: Q3_K_M is 8.1 GB, Q3_K_M is 8.1 GB, Q3_K_M is 8.1 GB, Q3_K_M is 8.1 GB, Q3_K_M is 8.1 GB, Q3_K_M is 8.6 GB and 25 more.

Where can I download Wan21 FastMix GGUF?

The files are on Hugging Face in the Zuntan/Wan21-FastMix repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

Can I run Wan21 FastMix on 8 GB VRAM?

Yes, the Q3_K_M quantization requires about 9.6 GB of VRAM at 8k context, which fits within an 8 GB GPU with some headroom.

Does the model support text prompts?

The model is focused on image‑to‑video generation and does not include explicit text prompt handling in the provided configuration.

Are there any licensing restrictions?

No license information is provided in the model card excerpt; users should verify usage rights before deployment.

Similar models

Browse all →

More from Zuntan

All by Zuntan →

More Wan models

All Wan →

Other medium models

All medium models →