whatmodel.sh / image / Step1X-Edit v1.2

Step1X-Edit v1.2

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

Step1X-Edit v1.2 is an open-weights image model from StepFun, released November 2025. Its published size is 12.44B parameters. Licence: apache-2.0.

Model information sources

· Weights licence ↗ · Parameter count ↗

Diffusion transformer only: HF Hub safetensors total for the official repo (pinned revision), 12,442,484,544 = the three transformer/ shards. The Qwen-VL text encoder and the VAE are not included. Observed 2026-10-07.

53.0±1.9#72 of 144 image modelsfrom 7 benchmark results across 6 sources

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

$ how it compares

The image models ranked either side of Step1X-Edit v1.2, out of 197 with a score, then more from StepFun.

#modelscorewhereRAMreleased
87LLaDA-Image-Turbo Ant Group54±3local~22GB2026-09
88LongCat-Image Meituan53±4api $0.13/image +local~15GB2025-12
89Step1X-Edit v1.2 StepFun53±2local~51GB2025-11
90Kling Image 2.1 Kuaishou53±7api $0.014/image
91Imagen 4 Fast Google53±4comparison only2025-08
more from StepFun
103Step1X-Edit v1.2 Preview StepFun50±5comparison only2025-09
108Step1X-Edit v1.1 StepFun48±3local~51GB2025-07
154NextStep-1-Large-Edit StepFun37±3comparison only2025-08

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

$ what it's good at

image to image / editing▁▁·▃▂▃█▃▃▄▁▁ #44 of 62 · 72% of its score · 5 tests

Beats GPT Image 1 Mini, FLUX.1 Kontext [max], Qwen-Image here, though they rank higher overall. Trails GLM-Image, FLUX.2 [klein] 4B here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
Edit-Compass (EN, overall)51.6% ✎▃▃▂·▃▆▂█·▃·▃ beats 12/20hard · avg 45%▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 41%at its level
KRIS-Bench overall lab60.9% ✎██·········█ beats 2/2medium · avg 54%▇▇▇ 8.6%at its level · small run
GEditBench v2 (overall Elo) lab996.0 ✎▃▂····▂█▂▃▂▃ beats 4/14avg 956▇▇▇ 8.5%blunt
GEdit-Bench overall (G_O, language split unspecified) lab75.8% ✎█·······█··█ beats 2/2easy · avg 73%▇▇▇ 8.0%small run
CREval-Bench creative image editing (Overall, GPT-4o QA evaluator)69.5% ✎▄█▄··█··▄·▄▄ beats 3/8easy · avg 74%▇▇▇ 6.3%blunt
other tests2 not in a task · 27% of its score
testresultvs the others in its rundifficultycountswhy
ChartSync chart editing (Overall)45.9% ✎▂··▃█▂▄▂▂·▂▂ beats 1/12medium · avg 41%▇▇▇▇▇▇▇▇ 19%at its level
WEAVEBench in-context multi-turn editing (Avg)60.5% ✎█▄▄·▄▄▄▄█·▄█ beats 8/12medium · avg 65%▇▇▇ 7.1%at its level

$ where to run it

Other ways to run it

  1. Run it as published · needs ~50.7 GB, NVIDIA only · from huggingface.co/stepfun-ai
    git clone -b step1xedit_v1p2 https://github.com/Peyton-Chen/diffusers.git && cd diffusers && pip install -e . && pip install transformers==4.55.0 && python -c "import torch; from diffusers import Step1XEditPipelineV1P2; from diffusers.utils import load_image; p=Step1XEditPipelineV1P2.from_pretrained('stepfun-ai/Step1X-Edit-v1p2', torch_dtype=torch.bfloat16); p.to('cuda'); o=p(image=load_image('input.png').convert('RGB'), prompt='add a ruby pendant on the neck of the girl.', num_inference_steps=50, true_cfg_scale=6, generator=torch.Generator().manual_seed(42), enable_thinking_mode=True, enable_reflection_mode=True); o.final_images[0].save('0001-final.jpg')"

Weights and documentation: huggingface.co/stepfun-ai ↗

$ how the score was worked out · Edit-Compass & EditReward-Com… 41%, ChartSync: A Benchmark for Vi… 19%, stepfun-ai/Step1X-Edit-v1p2 m… 17% and 3 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
Edit-Compass & EditReward-Compass: A Unified Benchmark for Image Editing and Reward Modeling1▇▇▇▇▇▇▇▇▇▇▇▇ 41%CC-BY-4.0
ChartSync: A Benchmark for Visuo-Logical Cascading Chart Editing (arXiv:2607.10301v1)1▇▇▇▇▇▇ 19%CC-BY-4.0
stepfun-ai/Step1X-Edit-v1p2 model card · lab-reported2▇▇▇▇▇ 17%Apache-2.0
GEditBench v2: A Human-Aligned Benchmark for General Image Editing · lab-reported1▇▇▇ 9%CC-BY-4.0
WEAVE: Unleashing and Benchmarking the In-context Interleaved Comprehension and Generation (arXiv:2511.11434v1)1▇▇ 7%CC-BY-4.0
CREval: An Automated Interpretable Evaluation for Creative Image Manipulation under Complex Instructions (arXiv:2603.26174v1)1▇▇ 6%CC-BY-4.0
prior · pull toward the middle▇ 2%
weighted average= 53.0

$ 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