Stability AI makes its latest Stable Diffusion 3 model available to developers via the Stability AI Developer Platform API.

The new Stable Diffusion 3 was released in preview two months ago and is now available to developers via the Stability AI Developer Platform API. This allows developers to try out the new image generation features. Additionally, the company also announced Stable Diffusion 3 Turbo, a faster version than Stable Diffusion 3 that will also be available to developers via an API.

For developers

Stability AI released the preview version of its new text-to-image AI model two months ago. Developers now have access to the new image generation functions via the API.

Stable Diffusion 3 is based on the Multimodal Diffusion Transformer (MMDiT) architecture, which uses separate sets of weights for visual and linguistic representations. This improves text comprehension and spelling options compared to previous versions of Stable Diffusion.

Coming soon

According to a research report published by Stability AI, this model based on human preference assessments is equivalent or better than state-of-the-art models such as DALL-E 3 and Midjourney v6. Although the model is available via an API, the company is continually making improvements leading up to the open release. Stability AI’s goal is to soon make the model available for self-hosting with a Stability AI membership.