video model · cogvideox · Windows
CV Can I run CogVideoX-5B on Nvidia GeForce RTX 3060 (12GB)?
Needs ~16 GB at INT8 / fp8, but only ~11 GB is usable on Nvidia GeForce RTX 3060 (12GB). With aggressive CPU offload it can run on as little as ~5 GB, much slower.
Needs ~16 GB at INT8 / fp8, but only ~11 GB is usable on Nvidia GeForce RTX 3060 (12GB). With aggressive CPU offload it can run on as little as ~5 GB, much slower.
- Peak VRAM
- ~16 GB
- Usable on device
- ~11 GB
- Device memory
- 12 GB
- Quant
- INT8 / fp8
- Type
- video (DIT)
- Parameters
- 5B
- Peak VRAM
- ~16 GB at INT8 / fp8
- Resolution
- 720×480
- License
- CogVideoX License
- Memory
- 12 GB vram
- Usable for weights
- ~11 GB
- Best runtime
- Ollama (CUDA) / llama.cpp CUDA
What you can run instead
Run CogVideoX-5B on other hardware
FAQ
Can Nvidia GeForce RTX 3060 (12GB) run CogVideoX-5B?
Needs ~16 GB at INT8 / fp8, but only ~11 GB is usable on Nvidia GeForce RTX 3060 (12GB). With aggressive CPU offload it can run on as little as ~5 GB, much slower.
How much VRAM does CogVideoX-5B need?
Nvidia GeForce RTX 3060 (12GB) does not have enough memory. At INT8 / fp8 the realistic peak is ~16 GB of VRAM, versus ~26 GB with every component kept resident (no offload). With aggressive CPU offload it drops to ~5 GB, much slower.
What do I use to run CogVideoX-5B locally?
CogVideoX-5B runs in Diffusers or ComfyUI. 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.