Magic Wan T2IV V3 needs about 12.6 GB of memory at Q5_K_M (a 10.8 GB file) with an 8K context window, so it fits on a 16 GB GPU. This repository has 3 quants, from 9.3 GB (Q4_1) to 15.4 GB (Q8_0).
14.3B parameter video generation model
Magic Wan T2IV V3 is a 14.3B parameter video generation model based on Wan2.2. It supports text‑to‑video and image‑to‑video generation with realistic scenes and facial detail. The model is released under Apache‑2.0 and is a pure base model suitable for fine‑tuning or LoRA training.
Good at
- Realistic video generation
- Unified T2I and T2V capabilities
- Supports English and limited Chinese font rendering
Limits
- Requires at least 10.9 GB VRAM for Q4_1
- No vision inference support
- No built‑in NSFW filtering
Best for
- Short‑length text‑to‑video projects
- Fine‑tuning for custom video styles
- LoRA training for specialized domains
Who it’s for. Developers and researchers needing a high‑quality, open‑source video generation model. The model can be used as a base for further training or combined with LoRA adapters for specific use cases.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q4_1 | 9.3 GB | 10.9 GB | Too big | Get file → | |
| Q5_K_M | 10.8 GB | 12.6 GB | Too big | Get file → | |
| Q8_0 | 15.4 GB | 17.8 GB | Too big | Get file → |
For 8 GB VRAM, the Q4_1 quant is the smallest at 9.27 GB file size and needs 10.9 GB VRAM. If you have 12 GB VRAM, Q5_K_M (10.8 GB, 12.6 GB VRAM) is a good choice. With 18 GB VRAM, Q8_0 (15.41 GB, 17.8 GB VRAM) offers the highest precision. Estimates are approximate.
What GPU do you need for Magic Wan T2IV V3?
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_1 (9.3 GB file, about 10.9 GB in use).
- 16 GB VRAM: the best quant that fits is Q5_K_M (10.8 GB file, about 12.6 GB in use).
- 24 GB VRAM: the best quant that fits is Q8_0 (15.4 GB file, about 17.8 GB in use).
Which Magic Wan T2IV V3 quant should you choose?
Q5_K_M (10.8 GB) is the middle option in this repository. Closer to the original model than 4-bit, for a moderate increase in size. A good pick when you have memory to spare.
If memory is tight, the smallest file is Q4_1 at 9.3 GB, with the quality trade-offs that come with fewer bits. The usual sweet spot between file size and quality, and the quant most people start with.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about Magic Wan T2IV V3
How much VRAM does Magic Wan T2IV V3 need?
About 12.6 GB at Q5_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 Magic Wan T2IV V3 on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 10.9 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Magic Wan T2IV V3 GGUF?
It depends on the quant: Q4_1 is 9.3 GB, Q5_K_M is 10.8 GB, Q8_0 is 15.4 GB.
Where can I download Magic Wan T2IV V3 GGUF?
The files are on Hugging Face in the wikeeyang/Magic-Wan-T2IV-V3 repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.
What license does Magic Wan T2IV V3 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 Magic Wan T2IV V3 on 8 GB VRAM?
No. The smallest quant (Q4_1) needs 10.9 GB VRAM at 8k context.
Does it support Chinese fonts?
Yes, it supports English and a limited set of Chinese fonts.
Is it fine‑tunable?
Yes, it is a pure base model suitable for fine‑tuning or LoRA training.