top of page

Stability AI launches SDXL 0.9: The Next Step in AI Art

Updated: Jun 29

Stability AI, a leading provider of artificial intelligence (AI) solutions, has announced a groundbreaking development in their Stable Diffusion text-to-image suite of models. SDXL 0.9, the latest iteration of this suite, represents a significant leap forward in generating highly detailed and realistic images compared to its predecessor.

stability ai art girl

A Step Forward for Stability AI

Building on the success of the beta release of Stable Diffusion XL in April, SDXL 0.9 introduces a host of improvements in image quality and composition. This advancement allows creators to generate hyper-realistic visuals for various applications such as films, television, music, instructional videos, design, and industrial use. By pushing the boundaries of generative AI imagery, SDXL emerges as a frontrunner in real-world applications for AI-generated visuals.

stability ai art aliens

Image to Image Functionality

One of the standout features of the SDXL series is its versatility, offering functionalities that extend beyond basic text prompting. For instance, the suite enables image-to-image prompting, allowing users to input one image and obtain variations of that image. Additionally, SDXL facilitates inpainting, which involves reconstructing missing parts of an image, and outpainting, which seamlessly extends existing images.

stability ai art wolves

SDXL 0.9 Pushing the Boundaries

The driving force behind the compositional advancements in Stability AI's SDXL 0.9 lies in its substantially increased parameter count. Parameters refer to the weights and biases within the neural network that the model is trained on. SDXL 0.9 boasts one of the largest parameter counts among open-source image models, with a base model of 3.5 billion parameters and a model ensemble pipeline consisting of 6.6 billion parameters. The ensemble pipeline aggregates the results of running the input through two models, with the second stage model adding finer details to the output generated by the first stage.

In comparison, the beta version of SDXL runs on 3.1 billion parameters and utilizes a single model. The notable increase in parameter count empowers SDXL 0.9 to produce imagery with enhanced depth and higher resolution. The model's processing power is further augmented by running on two CLIP models, including OpenCLIP ViT-G/14, which is one of the largest OpenCLIP models trained to date. With a resolution of 1024x1024, SDXL 0.9 pushes the boundaries of visual fidelity.

stability ai art bottle
Prompt: beautiful scenery nature glass bottle landscape, purple galaxy bottle (SDXL 0.9 - 1024x1024)

What to Look Forward To

To delve deeper into the specifications and testing of this revolutionary model, the Stability AI SDXL team plans to release a research blog in the near future. This resource will provide comprehensive insights into the workings of SDXL 0.9, offering a valuable resource for researchers, developers, and AI enthusiasts.

Highlights of SDXL’s capabilities include:

  • Next-level photorealism capabilities

  • Enhanced image composition and face generation

  • Rich visuals and jaw-dropping aesthetics

  • Use of shorter prompts to create descriptive imagery

  • Greater capability to produce legible text

As SDXL 0.9 ushers in a new era of AI-generated imagery, Stability AI is actively working on making the model accessible to users. The model can already be accessed via ClipDrop, while the API is expected to be made available shortly. Furthermore, an open release is scheduled for mid-July, marking the transition to SDXL 1.0 and ensuring wider availability for users to leverage the creative potential of this state-of-the-art AI technology.

With its remarkable advancements in composition, parameter count, and processing power, SDXL 0.9 from Stability AI has established itself as a game-changer in the field of generative AI imagery. This breakthrough has opened up a myriad of possibilities for various industries, empowering creators to bring their imaginative ideas to life with unparalleled realism and detail.

ai art kingdom banner stability ai

If you'd like to know more you can head over to for a curated collection of today's most popular, most liked AI artwork from across the internet. Plus explore an extensive array of AI tools, complemented by comprehensive guides and reviews, on our AI blog.


leonardo ai art promo
affiliate disclaimer
bottom of page