HunyuanVideo 1.5 in FP8 needs 18.5 GB of GPU memory or more (1280×720, 121 frames). The cheapest option right now is RTX 3090 24 GB, from $0.308/h.
“Minimum” is what ComfyUI does by default when memory is short. “No swapping” keeps everything loaded, so repeated runs are faster.
| Weights | Model | Text encoder | Minimum | No swapping | Fits on |
|---|---|---|---|---|---|
| BF16 | 15.5 | 15.4 | 26.2 | 41.7 | Tesla V100 32 GB |
| FP8 | 7.8 | 8.7 | 18.5 | 27.2 | RTX 4000 Ada 20 GB |
| GGUF Q8 | 8.2 | 8.7 | 18.9 | 27.7 | RTX 4000 Ada 20 GB |
| GGUF Q4 | 4.3 | 8.7 | 15.1 | 23.8 | RTX A4000 16 GB |
Working memory is our estimate for the listed output size and frame count. ComfyUI also runs the model on a smaller card by swapping part of the weights to system RAM, only slower.
File sizes checked against Hugging Face on 2026-10-19. 1 GB here is 1024³ bytes, like GPU memory. Hugging Face lists decimal gigabytes, so its numbers are 7% larger.
GPUs from the catalog the model fits on, by price per hour.
The price is the cheapest GPU of the model × the GPU count. A 2–8 GPU server can cost a little more per GPU.
The “ComfyUI” template starts the model in a ready environment: open the link once the server has loaded.
In ComfyUI the model goes into the models folder; the template ships with ComfyUI Manager.