April 25th.Adobe A blog post was published yesterday (April 24) launching the Firefly Image Model 4 and Firefly Image Model 4 Ultra, two AI models for text-generated imagesand previews updates to Creative Cloud apps for Photoshop and Illustrator.
Emphasizing speed and control, Firefly Image Model 4 has been called by Adobe "the fastest, most controllable, and most realistic image model to date," with support for generating images up to 2K resolution, with more precise control over style, size, and camera angle.
According to a blog post cited by 1AI, Firefly Image Model 4 Ultra specializes in "complex scenes and small structures," bringing a higher level of detail and realism. Both models ensure a fast and efficient generation process while significantly improving image quality over their predecessors. Both models are now available through the Firefly web app.
In addition to image modeling, Adobe has also launched Firefly Boards, a collaborative AI mood boarding tool similar to FigJam, which is currently in public beta.
In addition, the Firefly web app adds support for third-party AI models, allowing users to choose OpenAI's GPT image model, Google's Imagen 3 and Veo 2 models for image and video generation, and in the future will also support Luma, Pika, Runway, and other models.
Adobe has also brought several updates to Creative Cloud apps, including Illustrator's generative shape fills and text-to-pattern tools that help designers create more efficiently and effectively, which were tested last year, and Photoshop, which has optimized color adjustments and added the ability to automatically select hair, clothing, and facial features, as well as updated Actions panel, which offers smart editing suggestions based on a creator's unique style.