whatmodel.sh / video / LTX-2

LTX-2

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

LTX-2 is an open-weights video model from Lightricks, released October 2025. Its published size is 19B parameters. Licence: ltx-2-community.

49.0±1.8#53 of 93 video modelsfrom 4 benchmark results across 3 sources

$ how it compares

The video models ranked either side of LTX-2, out of 139 with a score, then more from Lightricks.

#modelscorewhereRAMreleased
68Runway Gen-4 Runway50±4comparison only2025-03
69Veo 2 Google50±2comparison only2025-04
70LTX-2 Lightricks49±2local~120GB2025-10
71Open-Sora Plan v1.3 (T2V) Peking University (PKU-YuanGroup)49±7local~24GB2024-10
72Cosmos-1.0-Diffusion-7B-Video2World NVIDIA49±5comparison only2025-01
more from Lightricks
53LTX-2.5 Lightricks53±6api $0.12/sec +local~14GB2026-08
83LTX-2 19B Distilled Lightricks47±8api $0.020/sec +local~41GB2026-01
112LTX-2.3 Lightricks36±3api $0.040/sec +local~34GB2026-03

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

$ what it's good at

text to video▆▄▃▄▆▅█▃▂▂▂▄ #34 of 68 * · 61% of its score · 2 tests

Beats Wan 2.2 A14B here, though they rank higher overall.

testresultvs the others in its rundifficultycountswhy
LMArena video1154.1 ✎▂▂▃▆█▆·▆▅▁▂▇ beats 7/41avg 1147▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 53%typical
WorldJen human preference (Bradley-Terry)1479.0 ✎█···██·█··██ beats 2/5avg 1470▇▇▇ 7.6%typical
image to video▁▂·▃▅█▅▄▅▆▃▄ #47 of 50 * · 35% of its score · 1 test

Trails Wan 2.1 I2V Turbo, HunyuanVideo-I2V, Hailuo 01 (MiniMax video-01 / I2V-01) here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
Arena image-to-video1159.4 ✎▁▁▁▃▄▇▃▆▃▂▄█ beats 3/40avg 1189▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 35%typical
other tests1 not in a task · 3.3% of its score
testresultvs the others in its rundifficultycountswhy
PhyGround physics human evaluation (Overall, TI2V)51.2% ✎█▄▄··▄···▄·█ beats 0/7medium · avg 57%▇ 3.3%blunt · at its level

$ where to run it

Other ways to run it

  1. Run it as published · needs ~119.6 GB, Mac only · from github.com/Blaizzy
    pip install git+https://github.com/Blaizzy/mlx-video.git@87db56a51758fefb748a359b90a5283bb8ba4837 && mlx_video.ltx_2.generate --pipeline dev-two-stage-hq --prompt "A cinematic scene of ocean waves at golden hour" --model-repo prince-canuma/LTX-2-dev

Official weights and documentation: huggingface.co/Lightricks ↗

$ how the score was worked out · LMArena leaderboard dataset 88%, WorldJen: An End-to-End Multi… 8%, PhyGround: Benchmarking Physi… 3%

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
LMArena leaderboard dataset2▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 88%CC BY 4.0
WorldJen: An End-to-End Multi-Dimensional Benchmark for Generative Video Models1▇▇ 8%CC-BY-4.0
PhyGround: Benchmarking Physical Reasoning in Generative World Models (arXiv:2605.10806v1)1▇ 3%CC-BY-4.0
prior · pull toward the middle▇ 1%
weighted average= 49.0

$ 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