Adobe announced the launch of a new version of its Firefly AI model for generating images, as well as an algorithm for generating vector graphics and an updated web application that brings together all of the company’s generative models, as well as some of its competitors’ neural networks. In addition, the developers continue to work on the Firefly mobile app.
Image source: Rubaitul Azad / Unsplash
The Firefly Image Model 4 large language model, according to Adobe, outperforms its predecessors in the quality of generated images, the speed of query processing, and the ability to customize the parameters of the created content. It supports the generation of images with a resolution of up to 2K. There is also a more powerful version of the Image Model 4 Ultra algorithm, which can create complex scenes with many small structures and a large number of details.
An Adobe representative said that the developers made the new AI models more productive so that they could generate more detailed images. Among other things, the generation of text on images has become more qualitative, and it is also possible to create several images in the same style as the original.
At the same time, the company opened access to its AI video generator Firefly, which entered beta testing last year, for everyone. The algorithm allows you to create videos based on a text description or image, change camera angles, specify the start and end frames, customize animation style elements, etc. The AI model can create videos in 1080p format.
Firefly’s vector-based AI model can generate editable vector illustrations, as well as iterate and generate variations of logos, patterns, and more. All of Adobe’s new AI models can be accessed in the updated Firefly web app. It also includes OpenAI’s GPT image generator, Google’s Imagen 3 and Veo 2 models, and Flux’s Flux 1.1 Pro algorithm. Users can switch between these algorithms at their discretion.
Adobe is also publicly testing a new product called Firefly Boards, which is a canvas for creativity and ideas. It can be used to generate or import images, edit them, and collaborate with other users of the platform. Firefly Boards will also be available in the Firefly web app.
In addition, Adobe has opened up its Text-to-Image and Avatar APIs, and announced beta testing of its Text-to-Video API. These and other APIs can be accessed through the company’s Firefly Services platform.