All tasks / Image / Image generation
Best local AI for image generation
Open-weights models and tools that run entirely on your own hardware, ranked from public benchmarks. Local tools are rated by the best open model they run.
BTop ranked: Qwen-Image-2.1 (Alibaba (Qwen)), 1,042 on Artificial Analysis Text to Image Arena (AA-Image-T2I v2.1).26 options listed · data verified 19 Oct 2026
Ranked list
| Tier | Name | Approach | Runs | Rated by | Price | Skill |
|---|---|---|---|---|---|---|
| B |  Qwen-Image-2.1Alibaba (Qwen) | Models | Local | 1,042 on Artificial Analysis Text to Image Arena (AA-Image-T2I v2.1) | Open weights | — |
| B |  ComfyUIComfy Org | Workflow platforms | Local | via Qwen-Image-2.1 Cloud models: tier S | Free | advanced |
| B |  DiffusersHugging Face | Runtimes & APIs | Local | via Qwen-Image-2.1 | Free | advanced |
| B |  InvokeAIInvoke (open-source community) | Apps + Workflow platforms | Local | via Qwen-Image-2.1 Cloud models: tier S | Free | intermediate |
| B |  WanGP (Wan2GP)deepbeepmeep (community project) | Apps | Local | via Qwen-Image-2.1 | Free | intermediate |
| B |  stable-diffusion.cppleejet (community) | Runtimes & APIs | Local | via Qwen-Image-2.1 | Free | advanced |
| B | SD.Next vladmandic (community) | Apps | Local | via Qwen-Image-2.1 | Free | intermediate |
| B |  Stable Diffusion WebUI Forge NeoHaoming02 (community) | Apps | Local | via Qwen-Image-2.1 | Free | intermediate |
| B |  AmuseTensorStack | Apps | Local | via Qwen-Image-2.1 | Free | beginner |
| C |  Ideogram 4.0Ideogram | Models | Local | 1,026 on Artificial Analysis Text to Image Arena (AA-Image-T2I v2.1) | Open weights | — |
| C |  FLUX.2 [dev]Black Forest Labs | Models | Local | 1,000 on Artificial Analysis Text to Image Arena (AA-Image-T2I v2.1) | ~24 GB VRAM | — |
| C |  Ming-Image-0.1-DesigninclusionAI (Ant Group) | Models | Local | 999 on Artificial Analysis Text to Image Arena (AA-Image-T2I v2.1) | Open weights | — |
| C |  Cosmos3-Super-Text2ImageNVIDIA | Models | Local | 992 on Artificial Analysis Text to Image Arena (AA-Image-T2I v2.1) | Open weights | — |
| C |  HunyuanImage 3.0Tencent | Models | Local | 989 on Artificial Analysis Text to Image Arena (AA-Image-T2I v2.1) | Open weights | — |
| C |  HiDream-O1-ImageHiDream.ai | Models | Local | 974 on Artificial Analysis Text to Image Arena (AA-Image-T2I v2.1) | Open weights | — |
| C |  Z-Image TurboAlibaba (Tongyi-MAI) | Models | Local | 934 on Artificial Analysis Text to Image Arena (AA-Image-T2I v2.1) | ~16 GB VRAM | — |
| C |  ERNIE-Image-TurboBaidu | Models | Local | 914 on Artificial Analysis Text to Image Arena (AA-Image-T2I v2.1) | ~24 GB VRAM | — |
| C |  Krea 2 TurboKrea | Models | Local | 1,113 on LMArena Text-to-Image Arena | Open weights | — |
| C |  GLM-ImageZ.ai | Models | Local | 888 on Artificial Analysis Text to Image Arena (AA-Image-T2I v2.1) | Open weights | — |
| C |  LongCat-ImageMeituan (LongCat) | Models | Local | 858 on Artificial Analysis Text to Image Arena (AA-Image-T2I v2.1) | Open weights | — |
| C |  Stable Diffusion 3.5 LargeStability AI | Models | Local | 938 on LMArena Text-to-Image Arena | Open weights | — |
| C | Easy Diffusion Easy Diffusion (community) | Apps | Local | via FLUX.2 [dev] | Free | beginner |
| C |  SwarmUImcmonkeyprojects | Apps + Workflow platforms | Local | via FLUX.2 [dev] | Free | intermediate |
| C |  Krita AI DiffusionAcly (Interstice) | Apps | Local | via Z-Image Turbo | Free | intermediate |
| C |  Mochi DiffusionMochi Diffusion (community) | Apps | Local | via Stable Diffusion 3.5 Large | Free | beginner |
| C |  FastSD CPUrupeshs (community) | Apps | Local | via Stable Diffusion 3.5 Large | Free | intermediate |
Tiers come only from public benchmarks (Artificial Analysis Text to Image Arena (AA-Image-T2I v2.1), LMArena Text-to-Image Arena); tools without a result of their own inherit the tier of the best model they offer. How we rank.