CogVideoX-5B
Z.ai / Tsinghua KEG · zai-org/CogVideoX-5b
Weights license: Publisher terms: CogVideoX License · code and components may use other terms
Quality
No comparable score
No public, reusable benchmark result covers this model.
Reported by the publisher
Runs on
Enough memory 71
A publisher documents a path within this memory, on this platform (CUDA or Apple MPS). Conditional paths need offload or quantization.
Show all 71
One memory tier short 1
The same machines with the next memory size up would have enough.
Hardware profiles grouped by the publisher's documented minimum memory path and platform. Not a measured run.
Consider CogVideoX-5B from Z.ai / Tsinghua KEG for the text-to-video task (5B parameters). Published single-GPU footprints: SAT BF16 26 GB; optimized Diffusers BF16 from 5 GB, or INT8 (torchao) from 4.4 GB.
Diffusers figures use CPU offload, VAE slicing and tiling, tested on A100/H100. Published video configuration: 720×480, 6 seconds at 8 fps. These are configuration-specific results, not guaranteed minima for other GPUs.
Measured speed
No published measurement for this checkpoint
Generation time per clip is shown only when a source publishes it with hardware, build and setup. It is never estimated.
How to run
Published single-GPU footprints: SAT BF16 26 GB; optimized Diffusers BF16 from 5 GB, or INT8 (torchao) from 4.4 GB. Diffusers figures use CPU offload, VAE slicing and tiling, tested on A100/H100. Published video configuration: 720×480, 6 seconds at 8 fps. These are configuration-specific results, not guaranteed minima for other GPUs. Official hardware evidence
Documented ways to run it
Copied as printed in the linked source (read 2026-10-18); not run by this site.
# diffusers>=0.30.1
# transformers>=4.44.2
# accelerate>=0.33.0 (suggest install from source)
# imageio-ffmpeg>=0.5.1
pip install --upgrade transformers accelerate diffusers imageio-ffmpeg import torch
from diffusers import CogVideoXPipeline
from diffusers.utils import export_to_video
prompt = "A panda, dressed in a small, red jacket and a tiny hat, sits on a wooden stool in a serene bamboo forest. The panda's fluffy paws strum a miniature acoustic guitar, producing soft, melodic tunes. Nearby, a few other pandas gather, watching curiously and some clapping in rhythm. Sunlight filters through the tall bamboo, casting a gentle glow on the scene. The panda's face is expressive, showing concentration and joy as it plays. The background includes a small, flowing stream and vibrant green foliage, enhancing the peaceful and magical atmosphere of this unique musical performance."
pipe = CogVideoXPipeline.from_pretrained(
"THUDM/CogVideoX-5b",
torch_dtype=torch.bfloat16
)
pipe.enable_model_cpu_offload()
pipe.vae.enable_tiling()
video = pipe(
prompt=prompt,
num_videos_per_prompt=1,
num_inference_steps=50,
num_frames=49,
guidance_scale=6,
generator=torch.Generator(device="cuda").manual_seed(42),
).frames[0]
export_to_video(video, "output.mp4", fps=8)Evidence level: Publisher documentation linked; no execution reproduced by this site.
Before downloading
- Check the exact checkpoint and license linked here.
- Confirm a compatible runtime and precision or quantization for your platform.
- Allow for context, batch size and any extra media components beyond the weights.
Technical details
What the source and the catalog store for this checkpoint. Missing fields are left out, never zero.
- Checkpoint / version
- zai-org/CogVideoX-5b
- Task and modality
- text-to-video · video
- Count and basis
- 5B · manufacturer model card
- Released
- THUDM/CogVideo README (GitHub), Project Updates · read 2026-10-09source
- Published repository size
- 21.5 GBNot runtime memory or a minimum download.files
- Published components
- 3D causal VAE · Expert transformer · Text encoder
- Runtime and precision
- No local execution verified.
- Commercial use
- Allowed with conditions Register for a commercial license; over 1M monthly visits needs a further license · CogVideoX License (zai-org/CogVideoX-5b LICENSE) · read 2026-10-18license text
The model license has its own conditions; follow the linked license rather than treating it as a generic open-source license.
Sources
Identity and published specifications come from the checkpoint page. Benchmark scores belong to the named release and do not measure speed on your computer.
Catalog reviewed: 2026-09-22.


















































