whatmodel.sh / video / VideoCrafter2

VideoCrafter2

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

VideoCrafter2 is an open-weights video model from Tencent AI Lab, released January 2024. Licence: apache-2.0.

34.1±1.7#80 of 93 video modelsfrom 7 benchmark results across 4 sources

$ how it compares

The video models ranked either side of VideoCrafter2, out of 139 with a score.

#modelscorewhereRAMreleased
116Allegro Rhymes AI35±2local~9.3GB2024-10
117Kling 1.0 Kuaishou (Kling AI)35±7comparison only2024-06
118VideoCrafter2 Tencent AI Lab34±2local~9.4GB2024-01
119Cosmos-1.0-Autoregressive-4B NVIDIA34±5comparison only2025-01
120Pika 1.0 Pika32±3comparison only2023-11

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

$ what it's good at

text to video▆▄▃▄▆▅█▃▂▂▂▄ #55 of 68 · 99% of its score · 7 tests
testresultvs the others in its rundifficultycountswhy
GenAI-Arena text-to-video (anonymous Elo)1029.2 ✎▃▃···▅▃·▃▅·█ beats 5/10avg 1229▇▇▇▇▇▇▇▇▇▇▇ 27%sharp
VideoPhy (human, SA=1 & PC=1)19.0% ✎▄▄▄·▄▄█····▄ beats 5/7medium · avg 48%▇▇▇▇▇▇▇▇▇ 22%sharp
AIGVE-60K human perception MOS (LOVE)48.1% ✎▄▄·█··█▄▄▄▄▄ beats 4/10easy · avg 73%▇▇▇▇▇▇▇ 17%at its level
VideoPhy-2 (human, joint SA & PC)10.5% ✎▄···▄·▄·█··▄ beats 0/5hard · avg 33%▇▇▇▇ 10%typical
AIGVE-60K human text-video correspondence MOS (LOVE)51.1% ✎▄▄·█▄█▄··▄▄▄ beats 7/10easy · avg 70%▇▇▇▇ 10%at its level
VideoPhy (auto-evaluator average)41.0% ✎▅·▃▅··▃█···▃ beats 7/9medium · avg 56%▇▇▇ 7.0%at its level
ChronoMagic-Bench-150 GPT4o-MTScore (time-lapse metamorphosis)57.4% ✎▄·▄█▄▄▄·▄·▄█ beats 7/10medium · avg 71%▇▇ 5.5%at its level

$ where to run it

Other ways to run it

  1. Run it as published · needs ~9.4 GB, NVIDIA only · from github.com/AILab-CVC (official)
    git clone https://github.com/AILab-CVC/VideoCrafter && cd VideoCrafter && hf download VideoCrafter/VideoCrafter2 model.ckpt --local-dir checkpoints/base_512_v2 && sh scripts/run_text2video.sh

Weights and documentation: huggingface.co/VideoCrafter ↗

$ how the score was worked out · VideoPhy / VideoPhy-2 officia… 39%, GenAI-Arena (TIGER-Lab) text-… 27%, LOVE: Benchmarking and Evalua… 27% and 1 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▇▇▇▇▇▇▇▇▇▇▇▇ 39%MIT
GenAI-Arena (TIGER-Lab) text-to-video leaderboard1▇▇▇▇▇▇▇▇ 27%MIT
LOVE: Benchmarking and Evaluating Text-to-Video Generation and Video-to-Text Interpretation (AIGVE-60K)2▇▇▇▇▇▇▇▇ 27%CC-BY-4.0
ChronoMagic-Bench: A Benchmark for Metamorphic Evaluation of Text-to-Time-lapse Video Generation1▇▇ 6%CC-BY-4.0
prior · pull toward the middle▇ 1%
weighted average= 34.1

$ 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