Nunchaku Flux Dev + Kontext
Optimized Flux Dev and Flux Kontext workflows using MIT Han Lab's Nunchaku framework — up to 2× faster generation while keeping quality, with 768×768 in under 45 seconds on a 6GB RTX 4050. Includes FP8 and GGUF Kontext workflows.
✓ 100% free — download the workflow and run it locally.Preview
Requirements
- GPU: NVIDIA RTX 30XX/40XX/50XX (FP16)
- VRAM: 4–6 GB minimum
- OS: Windows
- Storage: 40 GB+ free
Model downloads
- Nunchaku Flux Kontext https://huggingface.co/mit-han-lab/nunchaku-flux.1-kontext-dev/tree/main
- Nunchaku Flux Dev https://huggingface.co/mit-han-lab/nunchaku-flux.1-dev/tree/main
- CLIP text encoders https://huggingface.co/comfyanonymous/flux_text_encoders/tree/main
- 1x-SwatKatsLite + 2xLexicaRRDBNet_Sharp (upscalers/VAE) https://huggingface.co/Thelocallab/2xLexicaRRDBNet_Sharp/tree/main
Custom nodes
- ComfyUI Manager https://github.com/ltdrdata/ComfyUI-Manager
- ComfyUI-Detail-Daemon https://github.com/Jonseed/ComfyUI-Detail-Daemon
- ComfyUI-nunchaku https://github.com/nunchaku-tech/ComfyUI-nunchaku
Setup steps
- Install ComfyUI Manager + its requirements, then the nunchaku and Detail-Daemon nodes.
- Download all models to their ComfyUI folders.
- Load a workflow below (Kontext FP8 or GGUF).
- Enter a detailed prompt (LLM enhancement recommended) and generate.
Notes & tips
- ~45s at 768×768 on a 6GB RTX 4050.
- Prefer one-click? A Nunchaku Flux installer is available.
Want the 1-click version?
Skip the manual model downloads and node installs — get a ready-to-run installer, or unlock all 75+ with Local Lab Pro.