Canva
Dream Lab
Canva now has a new image generator on their site. I tested it with a text prompt and got results comparable to DALL-E or Midjourney.
Stability AI
Stable Diffusion 3.5
Stability AI has released new versions of their image generation models, Stable Diffusion 3.5. The Large version has 8.1B parameters, Large Turbo is a distilled variant of Large with faster generation times and Medium is at 2.5B parameters. Stability emphasises that the models excel in customisability, performance, diversity of output and versatility of styles.
I tested the model through deepinfra.com and got reasonable results.
Apple
Apple Intelligence
Apple has released new versions of their Apple Intelligence capabilities with the Developer Beta release of iOS 18.2 and MacOS 15.2.
Siri ChatGPT integration has been improved. One can now use an existing user subscription for ChatGPT. ChatGPT subscription can also be updated within Apple’s operating system settings. The system also provides information about daily limits for ChatGPT use based on the level of subscription.
Apple has also introduced Image Playground to generate images in Animation and Illustration styles, using prompts, existing photos and style suggestions. A similar tool is Image Wand, which can generate images based on text in Notes and using the Animation, Illustration and Sketch styles.
You can find out more about these new features in this post and the two follow-ups.
Apple Intelligence Developer Libraries
Apple has released new developer libraries to allow developers to use the new Apple Intelligence tools. It is now possible to use Writing Tools (which allows a user to revise text, give it one of the predefined styles or tones) or Image Playground (which can be used to create images in one of the Animation or Illustration styles based on textual prompts and existing images).
It is also possible to introduce new application functionality to Siri so as to make use of Apple Intelligence features of Siri.
Flux.ai
Flux 1.1 Ultra
Flux.ai has released a new Ultra mode for their 1.1 version. This mode creates better images without compromising on speed.
Flux.1 Tools
Flux.ai has released the Flux.1 Tools suite, a set of tools to fine-tune image generation.
Flux.1 Fill covers pinpointing and outpointing. Flux.1 Depth uses a depth map generated from the image to enable structural guidance. Flux.1 Canny uses canny edges from the image to enable structural guidance. Flux.1 Redux allows mixing and recreating input images and text prompts.