Flux Family: Open Source Image Generation Models for ComfyUI
ComfyUI Wiki
Flux is an open-source image generation model series by Black Forest Labs. Three generations: FLUX.1 (3.8B DiT), FLUX.2 (32B dense transformer), and FLUX 3 (multimodal flow model for video, image, audio, and action).
F
Flux by Black Forest Labs
Text-to-ImageImage EditingOpen SourceOpen-source image generation model series by Black Forest Labs (founded by ex-Stable Diffusion team). Three generations spanning from 3.8B DiT to 32B dense transformer and multimodal FLUX 3, covering text-to-image, image editing, video-audio generation, and action prediction.
Generations
| Generation | Params | Architecture | Released | Highlights |
|---|---|---|---|---|
| FLUX.1 | 3.8B | Rectified Flow DiT | 2024-08 | Open weights, Fill / Redux / Canny / Depth / Kontext tools |
| FLUX.2 | 32B | Dense Transformer | 2025-11 | T2I + multi-ref editing, up to 4MP, 24GB+ VRAM (FP8) |
| FLUX 3 | Multimodal | Flow Matching (Self-Flow) | 2026-07 | Video+Audio, Image, Action; first multimodal FLUX |
Select a generation for tutorials, custom nodes, and related news. Curated weight downloads live in the Model Hub catalog.
Comments
Sign in with GitHub to join the discussion.