StabilityAI has officially launched Stable Diffusion 3.5, a new family of image generation models that promises significant advancements compared to the previous version, Stable Diffusion 3. The company has highlighted improvements in realism, faster adherence, and better text rendering, aspects highly requested by the community.
Subscribe to the Softonic newsletter and get the latest in tech, gaming, entertainment and deals right in your inbox.
Subscribe (it's FREE) ►Like its predecessor, SD3.5 comes in three sizes: large (Large, 8B), large turbo (Large Turbo, 8B), and medium (Medium, 2.6B). All these models are designed to be highly customizable and to run on hardware accessible to the average consumer. Additionally, they remain under the StabilityAI community license, which means they are available for general and non-commercial use for free.
The company admitted in a statement that the Stable Diffusion 3 Medium model, launched in June, “did not fully meet our standards or the expectations of our communities.” Therefore, StabilityAI decided to take its time to develop SD3.5 and advance in its mission to “transform visual media.”
One of the highlights of this new version is its ability to customize images with greater precision, allowing users to create more unique images. Additionally, the system has incorporated new stylistic options, such as photography or painting, and also accepts hashtags to adjust styles.
The company claims that the Stable Diffusion 3.5 Large model leads in terms of adherence to instructions and image quality, while Stable Diffusion 3.5 Large Turbo offers extremely fast inference times. Meanwhile, Stable Diffusion 3.5 Medium is the best option for those seeking a balance between performance and high visual quality.
Stable Diffusion 3.5 is free for non-commercial use, as well as for small and medium-sized businesses with revenues under one million dollars, but companies exceeding that amount will need to opt for an Enterprise license.
.png){kind=icon}
