How much VRAM does Wan 2.2 A14B need?
Wan 2.2 A14B needs about 20.4 GB at 8-bit, or 14.6 GB as a 4-bit GGUF, at its default output size.
- Type
- Video model
- Parameters
- 14.3B
- Text encoder
- UMT5-XXL
- Released
- Jul 2025
- License
- Apache 2.0
Memory at each precision
| Precision | Weights | Peak at 720p, 5 s |
|---|---|---|
| 16-bit BF16 | 26.6 GB | 33.2 GB |
| 8-bit FP8 · Q8_0 | 13.8 GB | 20.4 GB |
| 6-bit GGUF Q6_K | 11.0 GB | 17.6 GB |
| 5-bit GGUF Q5_K_M | 9.5 GB | 16.1 GB |
| 4-bit GGUF Q4_K_M · NVFP4 | 8.0 GB | 14.6 GB |
The UMT5-XXL text encoder (5.3 GB at FP8) runs first and then waits in system RAM.
What runs it
| Machine | Best fit | Per clip |
|---|---|---|
| RTX 3060 12 GB | Too big | |
| RTX 4060 Ti 16 GB | 4-bit | 2.2 h |
| RTX 3090 | 8-bit | 72 min |
| RTX 4090 | 8-bit | 31 min |
| RTX 5090 | 8-bit | 24 min |
| RTX PRO 6000 Blackwell | 16-bit | 10 min |
| M4 Max 128 GB | 16-bit | 5.5 h |
| M5 Max 128 GB | 16-bit | 1.6 h |
| Ryzen AI Max+ 395 128 GB | 16-bit | 2.4 h |
| DGX Spark 128 GB | 16-bit | 81 min |
| M3 Ultra 256 GB | 16-bit | 3.1 h |
| M5 Ultra 512 GB | 16-bit | 58 min |
| H100 SXM | 16-bit | 5.4 min |
| DGX Station GB300 | 16-bit | 2.3 min |
720p, 5 s, 40 steps. Times are rough, plus or minus 30%.
Where to get it
Questions
How much VRAM does Wan 2.2 A14B need?
Wan 2.2 A14B needs about 20.4 GB at 8-bit, or 14.6 GB as a 4-bit GGUF, at its default output size.
Can Wan 2.2 A14B run on an RTX 4090?
Yes. At 8-bit it takes about 31 min per clip.
More models
Data checked Oct 2026.