Best local image generation models
Diffusion models you can run yourself. Their published parameter count describes the denoiser alone; the runnable pipeline is larger.
From the file· live filter over real dataFrom the file· 40 models
How this is ranked
Ranked by downloads. Sizes shown are the smallest published quantization of the denoiser — see each model page for the full component graph.
Best local image generation models
| # | Model | Params○ | Smallest quant● | Smallest quant● |
|---|---|---|---|---|
| 1 | FLUX.2-klein-9B | 9.1B | 3.71 GiB | 3.71 GiB |
| 2 | Qwen-Image-Edit-2511 | 20.4B | 4.13 GiB | 4.13 GiB |
| 3 | FLUX.1-dev | 11.9B | 3.75 GiB | 3.75 GiB |
| 4 | Krea-2-Turbo | 12.8B | 2.78 GiB | 2.78 GiB |
| 5 | FLUX.2-klein-4B | 3.9B | 1.70 GiB | 1.70 GiB |
| 6 | Z-Image-Turbo | 6.2B | 2.41 GiB | 2.41 GiB |
| 7 | Chroma | 8.9B | 3.99 GiB | 3.99 GiB |
| 8 | FLUX.1-schnell | 11.9B | 3.73 GiB | 3.73 GiB |
| 9 | FLUX.2-dev | 32.2B | 11.98 GiB | 11.98 GiB |
| 10 | Qwen-Image-2512 | 20.4B | 6.83 GiB | 6.83 GiB |
| 11 | FLUX.1-Fill-dev | 11.9B | 3.87 GiB | 3.87 GiB |
| 12 | Qwen-Image-Edit-2509 | 20.4B | 6.50 GiB | 6.50 GiB |
| 13 | Qwen-Image | 20.4B | 6.58 GiB | 6.58 GiB |
| 14 | Krea-2-Raw | 12.8B | 2.34 GiB | 2.34 GiB |
| 15 | wai-nsfw-illustrious-v110-sdxl | 2.6B | 1.39 GiB | 1.39 GiB |
| 16 | stable-diffusion-3.5-large | 8.1B | 4.44 GiB | 4.44 GiB |
| 17 | FLUX.1-Kontext-dev | 11.9B | 3.75 GiB | 3.75 GiB |
| 18 | FLUX.2-klein-base-9B | 9.1B | 3.71 GiB | 3.71 GiB |
| 19 | Chroma1-HD | 8.9B | 3.18 GiB | 3.18 GiB |
| 20 | Z-Image | 6.2B | 3.74 GiB | 3.74 GiB |
| 21 | stable-diffusion-3.5-medium | 2.5B | 1.35 GiB | 1.35 GiB |
| 22 | stable-diffusion-v1-5 | 860M | 1.46 GiB | 1.46 GiB |
| 23 | Qwen-Image-Edit | 20.4B | 6.55 GiB | 6.55 GiB |
| 24 | ideogram-4-fp8 | 9.3B | 4.46 GiB | 4.46 GiB |
| 25 | flux2-klein-9b-uncensored-text-encoder | 8.2B | 3.06 GiB | 3.06 GiB |
| 26 | FLUX.1-Krea-dev | 11.9B | 3.66 GiB | 3.66 GiB |
| 27 | stable-diffusion-3-medium | 7.9B | 1.20 GiB | 1.20 GiB |
| 28 | Anima | 2.1B | 0.95 GiB | 0.95 GiB |
| 29 | ERNIE-Image-Turbo | 8.0B | 2.96 GiB | 2.96 GiB |
| 30 | stable-diffusion-xl-base-1.0 | 2.6B | 1.39 GiB | 1.39 GiB |
| 31 | HiDream-I1-Full | 17.1B | 6.11 GiB | 6.11 GiB |
| 32 | FLUX.2-klein-9b-kv | 9.1B | 3.33 GiB | 3.33 GiB |
| 33 | JoyAI-Image-Edit-Plus-Diffusers | 16.3B | 7.33 GiB | 7.33 GiB |
| 34 | Illustrious-xl-early-release-v0 | 2.6B | 3.30 GiB | 3.30 GiB |
| 35 | Pony_Diffusion_V6_XL | 2.6B | 2.61 GiB | 2.61 GiB |
| 36 | ERNIE-Image | 8.0B | 2.00 GiB | 2.00 GiB |
| 37 | HunyuanImage-2.1 | 17.5B | 5.26 GiB | 5.26 GiB |
| 38 | Boogu-Image-0.1-Turbo | 10.3B | 12.69 GiB | 12.69 GiB |
| 39 | Chroma1-Flash | 8.9B | 3.18 GiB | 3.18 GiB |
| 40 | FLUX.2-klein-base-4B | 3.9B | 1.70 GiB | 1.70 GiB |
Best local models for codingBest local reasoning modelsBest models for an 8GB cardBest models for long contextBest permissively licensed modelsBest models under 4GBLargest models with a runnable quantizationBest local speech recognition modelsBest local text-to-speech modelsBest local vision-language models