whatmodel.sh / image / VILA-U 7B (256px)

VILA-U 7B (256px)

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

VILA-U 7B (256px) is an open-weights image model from MIT Han Lab / NVIDIA, released October 2024. Its published size is 8.12B parameters. Licence: mit.

Model information sources

Release date ↗ · Weights licence ↗ · Parameter count ↗

OUR SUM of the safetensors headers of all official weights (pinned revision): LLM shards 6,738,448,384 + mm_projector 20,979,712 + vision_tower (unified RQ-VAE vision tower) 1,356,296,327 = 8,115,724,423. The name's '7B' refers to the LLM. Observed 2026-10-07.

32.1±6.5#128 of 144 image modelsfrom 1 benchmark result across 1 source* thinly tested, treat the rank as provisional

Score within the LMArena text-to-image comparison set only.

$ how it compares

The image models ranked either side of VILA-U 7B (256px), out of 197 with a score.

#modelscorewhereRAMreleased
172Lumina-DiMOO Alpha-VLLM (Shanghai AI Lab)32±4local~39GB2025-09
173UltraEdit (SD3, with mask) Peking University / BIGAI (UltraEdit)32±7comparison only2024-06
174VILA-U 7B (256px) MIT Han Lab / NVIDIA32±7local~20GB2024-10
175LCM-SSD-1B latent-consistency32±4local~3.6GB2023-11
176Janus-1.3B DeepSeek32±4local~5.5GB2024-10

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

$ what it's good at

text to image▂▅▅▃▄▄█▃▁▂▁▁ #115 of 126 * · 82% of its score · 1 test

Beats Latent Diffusion text2img-large (LDM-KL-8-G, LAION-400M), VQ-Diffusion (ITHQ) here, though they rank higher overall. Trails CogView2, Versatile Diffusion v1.0, Dreamlike Photoreal 2.0 here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
WISE (WiScore overall, GPT-4o judge)31.0% ✎▄·▆▂▆█▄▆▂▄·▄ beats 2/22medium · avg 60%▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 82%at its level

$ where to run it

Other ways to run it

  1. Run it as published · needs ~20 GB, NVIDIA only · from github.com/mit-han-lab (official)
    git clone https://github.com/mit-han-lab/vila-u && cd vila-u && ./environment_setup.sh vila-u && conda activate vila-u && git lfs install && git clone https://huggingface.co/mit-han-lab/vila-u-7b-256 && CUDA_VISIBLE_DEVICES=0 python inference.py --model_path vila-u-7b-256 --prompt "A snowy mountain." --save_path generated_images --generation_nums 8

Official weights and documentation: huggingface.co/mit-han-lab ↗

$ how the score was worked out · WISE: A World Knowledge-Infor… 82%

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
WISE: A World Knowledge-Informed Semantic Evaluation for Text-to-Image Generation (arXiv v4)1▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 82%CC-BY-SA-4.0
prior · pull toward the middle▇▇▇▇▇ 18%
weighted average= 32.1

$ 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