MODELS
Stable Diffusion 3.5 Large
Stability's open-weights image model — the workhorse for self-hosting and ComfyUI.
Specs
- Modalities
- image
- Tool use
- —
- Vision
- —
- Streaming
- —
- License
- stability-community
- Released
- 2024-10-22
Pricing
Stable Diffusion 3.5 Large (October 2024) is Stability AI's 8B-parameter image model with open weights under the Stability Community licence (free up to $1M/yr revenue). 1MP native resolution, MMDiT-X architecture, strong prompt fidelity. The default for ComfyUI / Automatic1111 / Forge users who want SOTA-class quality without API per-image fees. Lighter siblings: SD 3.5 Medium (2.5B), SD 3.5 Large Turbo (distilled, 4 steps).
Editor's verdict
Pick when self-hosting matters — privacy, no API rate limit, custom LoRA stack — and you're OK with output that's good but not quite FLUX.1 Pro level. The huge ComfyUI / LoRA / ControlNet ecosystem around SD is the real value: nothing else gives you this much control. For one-shot 'generate me a hero image' work, FLUX or Imagen are usually faster paths to a publishable result.
Reviews
No reviews yet. Be the first.
Last updated: 2026-04-29