whatmodel.sh / image / Stable Cascade

Stable Cascade

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

Stable Cascade is an open-weights image model from Stability AI, released February 2024. Its published size is 5.1B parameters. Licence: stable-cascade-nc-community.

Catalogue identity notes

stabilityai/stable-cascade + stable-cascade-prior (Wuerstchen v3). params_b 5.1 = 3.6B Stage C + 1.5B Stage B, the recommended large variants per the model card.

38.4±3.5#112 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 Stable Cascade, out of 197 with a score, then more from Stability AI.

#modelscorewhereRAMreleased
150Kolors Kuaishou (Kolors team)39±3local~22GB2024-07
151BLIP3-o 8B Salesforce Research38±3local~41GB2025-05
152Stable Cascade Stability AI38±4local~13GB2024-02
153Janus-Pro-7B DeepSeek38±2api $0.0020/image +local~18GB2025-01
154NextStep-1-Large-Edit StepFun37±3comparison only2025-08
more from Stability AI
137Stable Diffusion 3.5 Large Stability AI42±1api $0.065/image +local~5.4GB2024-10
141Stable Diffusion 3.5 Medium Stability AI41±3api $0.035/image +local~20GB2024-10
148Stable Diffusion XL 0.9 (base) Stability AI40±6local~4.9GB2023-06

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

$ what it's good at

text to image▂▅▅▃▄▄█▃▁▂▁▁ #92 of 126 * · 95% of its score · 1 test

Beats Latent Diffusion text2img-large (LDM-KL-8-G, LAION-400M), VQ-Diffusion (ITHQ), Stable Diffusion XL 0.9 (base) here, though they rank higher overall.

testresultvs the others in its rundifficultycountswhy
GenAI-Arena text-to-image (Elo, 4 Jun 2024)1057.0 ✎▃▃··▃·▃·█·▃▃ beats 6/8avg 1237▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 95%sharp

$ where to run it

Other ways to run it

  1. Run it as published · needs ~12.7 GB, NVIDIA only · from huggingface.co/stabilityai (official)
    uv run --with diffusers --with transformers --with accelerate --with torch python -c "import torch; from diffusers import StableCascadePriorPipeline, StableCascadeDecoderPipeline; pr=StableCascadePriorPipeline.from_pretrained('stabilityai/stable-cascade-prior', variant='bf16', torch_dtype=torch.bfloat16); de=StableCascadeDecoderPipeline.from_pretrained('stabilityai/stable-cascade', variant='bf16', torch_dtype=torch.float16); pr.enable_model_cpu_offload(); o=pr(prompt='a cat holding a sign', height=1024, width=1024, guidance_scale=4.0, num_inference_steps=20); de.enable_model_cpu_offload(); de(image_embeddings=o.image_embeddings.to(torch.float16), prompt='a cat holding a sign', guidance_scale=0.0, num_inference_steps=10, output_type='pil').images[0].save('out.png')"

Official weights and documentation: huggingface.co/stabilityai ↗

$ how the score was worked out · GenAI Arena: An Open Evaluati… 95%

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
GenAI Arena: An Open Evaluation Platform for Generative Models (arXiv v3)1▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 95%CC-BY-4.0
prior · pull toward the middle▇▇ 5%
weighted average= 38.4

$ 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