Stability AI has released a new version of its generative image model — Stable Diffusion 3.5. This version includes several model variants, including Stable Diffusion Large and Large Turbo, which are available for download now. Another model, Stable Diffusion Medium, will be released on October 29. All new models are free for non-commercial and commercial use, provided the user’s annual revenue does not exceed $1 million.
Stable Diffusion 3.5 Large uses 8 billion parameters and generates images with a resolution of up to 1 megapixel. The Large Turbo model offers a faster generation process while maintaining high quality. All models are designed to run on standard hardware, ensuring ease of use for any user.
During development, the company took into account community feedback on previous releases and made significant improvements to generation accuracy and model speed. According to Stability AI representatives, the 3.5 models provide greater diversity of results and can generate images with various styles and characteristics without the need for complex prompts.
You can download Stable Diffusion 3.5 from Hugging Face, and the source code is available on GitHub.
