Overview

Stable Diffusion 3.5, released in October 2024, is a family of text-to-image models from Stability AI. The Large version, with 8.1 billion parameters, uses the Multimodal Diffusion Transformer (MMDiT) architecture introduced with Stable Diffusion 3 and targets high-quality, prompt-faithful images at 1-megapixel resolution.

Highlights

  • 8.1 billion parameters, MMDiT architecture
  • Improved prompt adherence and typography over earlier versions
  • Turbo and Medium variants for speed and lower hardware needs
  • Broad ecosystem support in tools such as ComfyUI
  • Fine-tunable for custom styles

Typical uses

Image generation in creative workflows, product mock-ups, custom fine-tuned styles and self-hosted image services.

Things to consider

  • The Community License requires an enterprise license above an annual revenue threshold.
  • Needs a capable GPU for comfortable local use.
  • Review rights and safety policies for generated content.

Specifications are taken from the provider’s official announcement, documentation or model card. Benchmark results depend on evaluation settings; see how AI benchmarks work.