whatmodel.sh / video / CogVideoX-5B

CogVideoX-5B

+ contribute
Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

CogVideoX-5B is an open-weights video model from Zhipu AI (THUDM), released August 2024. Its published size is 5B parameters. Licence: cogvideox.

43.8±1.8#64 of 93 video modelsfrom 7 benchmark results across 5 sources

$ how it compares

The video models ranked either side of CogVideoX-5B, out of 139 with a score, then more from Zhipu AI (THUDM).

#modelscorewhereRAMreleased
89Step-Video-T2V StepFun45±7comparison only2025-02
90HunyuanVideo-I2V Tencent45±4local~57GB2025-03
91CogVideoX-5B Zhipu AI (THUDM)44±2api price unknown +local~13GB2024-08
92Step-Video-TI2V StepFun44±7comparison only2025-03
93Wan 2.1 I2V Turbo Alibaba43±7api price unknown2025-02
more from Zhipu AI (THUDM)
85CogVideoX-5B-I2V Zhipu AI (THUDM)46±4local~13GB2024-09
113CogVideoX1.5-5B Zhipu AI (THUDM)36±2local~38GB2024-11
114CogVideoX-2B Zhipu AI (THUDM)36±2local~5.3GB2024-08

see all video models · find the ones that fit your machine

$ what it's good at

text to video▆▄▃▄▆▅█▃▂▂▂▄ #42 of 68 · 87% of its score · 5 tests
testresultvs the others in its rundifficultycountswhy
VideoPhy (human, SA=1 & PC=1)39.6% ✎▄▄▄·▄▄█····▄ beats 7/7medium · avg 48%▇▇▇▇▇▇▇▇▇▇▇▇▇ 32%sharp · at its level
GenAI-Arena text-to-video (anonymous Elo)1125.6 ✎▃▃···▅▃·▃▅·█ beats 10/10avg 1229▇▇▇▇▇▇▇▇▇▇▇▇ 29%sharp
VideoPhy-2 (human, joint SA & PC)25.0% ✎▄···▄·▄·█··▄ beats 4/5hard · avg 33%▇▇▇▇▇▇ 16%typical
VideoPhy (auto-evaluator average)49.0% ✎▅·▃▅··▃█···▃ beats 9/9medium · avg 56%▇▇▇ 7.9%at its level
HumanScore (T2V)74.8% ✎▃··▂··█▄▂··▆ beats 0/11easy · avg 84%▇ 2.1%blunt
other tests2 not in a task · 11% of its score
testresultvs the others in its rundifficultycountswhy
VideoArgus-Bench T2V (rubric-grounded score)73.9% ✎█▄·▄█▄▄····▄ beats 0/8easy · avg 82%▇▇ 6.0%blunt
WorldModelBench (human total)7.31 / 10 ✎▅··▅▃···▃··█ beats 5/8easy · avg 7.71 / 10▇▇ 5.2%at its level

$ where to run it

Through an API

api · fal · price unknown · fal.ai ↗

Other ways to run it

  1. Run it as published · needs ~12.5 GB, NVIDIA only · from huggingface.co/THUDM (official)
    python -c "import torch;from diffusers import CogVideoXPipeline;from diffusers.utils import export_to_video;p=CogVideoXPipeline.from_pretrained('THUDM/CogVideoX-5b',torch_dtype=torch.bfloat16);p.enable_sequential_cpu_offload();p.vae.enable_slicing();p.vae.enable_tiling();export_to_video(p(prompt='A panda playing a tiny guitar in a bamboo forest',num_inference_steps=50,num_frames=49,guidance_scale=6).frames[0],'output.mp4',fps=8)"

Weights and documentation: huggingface.co/zai-org ↗

$ how the score was worked out · VideoPhy / VideoPhy-2 officia… 56%, GenAI-Arena (TIGER-Lab) text-… 29%, VideoArgus: Agentic Rubric-Gr… 6% and 2 more

Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works

sourceresultsshare of the scorelicence
VideoPhy / VideoPhy-2 official repository leaderboards3▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 56%MIT
GenAI-Arena (TIGER-Lab) text-to-video leaderboard1▇▇▇▇▇▇▇▇▇ 29%MIT
VideoArgus: Agentic Rubric-Grounded Unified Evaluation for Video Generation and Editing (arXiv:2608.05485v1)1▇▇ 6%CC-BY-4.0
WorldModelBench: Judging Video Generation Models As World Models1▇▇ 5%CC-BY-4.0
HumanScore: Benchmarking Human Motions in Generated Videos1▇ 2%CC-BY-4.0
prior · pull toward the middle▇ 1%
weighted average= 43.8

$ improve this page

Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.

data updated · sources and licences · json