Adobe on Thursday started the last repetition of the Firefly of Image Generation AI Models, a model for vectors and one redesigned web application This houses all AI models as well as some of its competitors. There is also a mobile app for Firefly in the projects.
The new Firefly 4 image model, says Adobe, improves its predecessors in terms of quality, speed and amount of control of the structure and style of expenses, camera and zoom corners. Can create images in resolutions up to 2K. There is also a modified, more capable version of this model called the 4 Ultra model image that can make complex scenes consisting of small structures and many details.
Alexandru Costin, Vice President of AI Genetics at Adobe, said the company is training models with a higher computational class to allow them to create more detailed images. He added that, compared to previous generation models, the new models improve the production of text in images and have features that allow users to use images of their choice to get the model to create images in this style.
The company also makes the Firefly video model, which started in limited beta last year, available to everyone. It allows users to create video clips from a line or text image, to use camera angles, to determine the starting and finishing frames to control the shots, to create atmospheric elements, and to customize motion design elements. The model can create a video clip from text to analyzes up to 1080p.
In the meantime, the Vector Firefly model can create works of art based on edited vector, repeat and create variations of logo, product packages, icons, scenes and patterns.
Specifically, the Firefly Web application gives you access to all these models as well as some models of image and videos from Openai (Gind Gerreration), Google (Imagen 3 and VEO 2) and Flow (Flux 1.1 Pro). Users can change between any of these models at any point and the images produced by any model will have the content certificates that are connected to them. The company hints that other AI models can be added to web application in the future.
Adobe also publicly tests a new product called Firefly Boards, a canvas for ideation or moodboarding. It allows users to create or introduce images, to reconstruct them and work with others-to what you can do with other AI-based ideas from Visual Electric, Cove or Kosmik. Boards are available through the Web Firefly app.
The company said these new models would soon be integrated into its product portfolio, but did not give a timetable for growth.
Adobe also makes API’s API and API API and stated that a new video API in video is now available for use in Beta. These APIs are available through the API Firefly collection, tools and services.
Adobe also tests a web application called Authenticity of Adobe Content To allow users to attach their credentials to indicate ownership and performance through metadata. Users can also indicate whether AI companies can use any images for AI models training.
