video model · cogvideox · Windows
Can I run CogVideoX-2B on Nvidia GeForce RTX 2060 (6GB)?
Needs ~8 GB at fp16 + offload, but only ~5 GB is usable on Nvidia GeForce RTX 2060 (6GB). With aggressive CPU offload it can run on as little as ~4 GB, much slower.
Needs ~8 GB at fp16 + offload, but only ~5 GB is usable on Nvidia GeForce RTX 2060 (6GB). With aggressive CPU offload it can run on as little as ~4 GB, much slower.
The gap is about 3 GB: CogVideoX-2B needs roughly 8 GB and Nvidia GeForce RTX 2060 (6GB) leaves only about 5 GB usable for a model. The lightest tracked hardware that runs CogVideoX-2B is the Nvidia GeForce RTX 3060 (12GB) at 12 GB. See CogVideoX-2B on Nvidia GeForce RTX 3060 (12GB).
- Peak VRAM
- ~8 GB
- Usable on device
- ~5 GB
- Device memory
- 6 GB
- Quant
- fp16 + offload
- Type
- video (DIT)
- Parameters
- 2B
- Peak VRAM
- ~8 GB at fp16 + offload
- Resolution
- 720×480
- License
- Apache-2.0
- Memory
- 6 GB vram
- Usable for weights
- ~5 GB
- Power draw
- ~160 W
- Best runtime
- Ollama (CUDA) / llama.cpp CUDA
Run CogVideoX-2B on other hardware
FAQ
Can Nvidia GeForce RTX 2060 (6GB) run CogVideoX-2B?
Needs ~8 GB at fp16 + offload, but only ~5 GB is usable on Nvidia GeForce RTX 2060 (6GB). With aggressive CPU offload it can run on as little as ~4 GB, much slower.
How much VRAM does CogVideoX-2B need?
Nvidia GeForce RTX 2060 (6GB) does not have enough memory. At fp16 + offload the realistic peak is ~8 GB of VRAM, versus ~18 GB with every component kept resident (no offload). With aggressive CPU offload it drops to ~4 GB, much slower.
What do I use to run CogVideoX-2B locally?
CogVideoX-2B 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 2026-09-14. See methodology.