ComfyUI
A node graph for image and video models. Steeper to learn than anything else here, and the only one that lets you build exactly the pipeline you want.
- Self-hosted
- Allowed
FLUX.1 [dev]
A large open-weights image model with notably better prompt adherence and text rendering than the Stable Diffusion line.
- Open weights
- Image
- Filtered adult content
At a glance
- Made by
- Black Forest Labs
- Family
- FLUX
- Licence
- FLUX.1 [dev] Non-Commercial License
- Graphics memory
- 12 GB minimum, 24 GB recommended
- Adult content
- FilteredA filter blocks it, with varying success.
Open weights, but not an open-source licence: the [dev] weights are non-commercial. It is a 12-billion-parameter model, so the memory requirement is real — quantised community builds bring it onto 8 and 12 GB cards at a quality cost.
The released weights are distilled from a larger model and are noticeably reluctant about anatomy; the community fine-tunes exist for the same reason they do for SDXL.