video model · wan · Windows
WA Can I run Wan 2.2 T2V A14B on Nvidia GeForce RTX 3090 (24GB)?
Yes. Wan 2.2 T2V A14B runs on Nvidia GeForce RTX 3090 (24GB) at Q4 GGUF (~16 GB of ~23 GB usable).
Runs at Q4 GGUF using ~16 GB of ~23 GB usable.
- Peak VRAM
- ~16 GB
- Usable on device
- ~23 GB
- Device memory
- 24 GB
- Quant
- Q4 GGUF
How to run it
Use ComfyUI or Diffusers at Q4 GGUF. The big text encoder is loaded to encode your prompt, then offloaded before generation, which is why peak VRAM stays near the backbone size rather than the sum of every file.
- Type
- video (DIT)
- Parameters
- 27B (MoE, 14B active)
- Peak VRAM
- ~16 GB at Q4 GGUF
- Resolution
- 1280×720 (720p)
- License
- Apache-2.0
- Memory
- 24 GB vram
- Usable for weights
- ~23 GB
- Best runtime
- vLLM (Linux) / Ollama (CUDA)
You could also run
Run Wan 2.2 T2V A14B on other hardware
FAQ
Can Nvidia GeForce RTX 3090 (24GB) run Wan 2.2 T2V A14B?
Yes. Wan 2.2 T2V A14B runs on Nvidia GeForce RTX 3090 (24GB) at Q4 GGUF (~16 GB of ~23 GB usable).
How much VRAM does Wan 2.2 T2V A14B need?
Nvidia GeForce RTX 3090 (24GB) has room to spare. At Q4 GGUF the realistic peak is ~16 GB of VRAM, versus ~80 GB with every component kept resident (no offload). With aggressive CPU offload it drops to ~8 GB, much slower.
What do I use to run Wan 2.2 T2V A14B locally?
Wan 2.2 T2V A14B runs in ComfyUI or Diffusers. It loads as a video diffusion checkpoint plus its text encoder and VAE, not a single chat command.
Sources
VRAM figures are sourced peak-usage anchors at the noted quant, validated 2026-06-15. See methodology.