Adobe unveiled today the next version of its Firefly generative AI models, which include the Adobe Firefly Image 2 Model, which dramatically improves Photoshop’s text-to-image feature, the new Adobe Firefly Vector Model that powers text-to-vector generation, and the Adobe Firefly Design Model to convert text into graphic templates.
All these models fuel multiple new and improved AI features within Adobe Photoshop, Illustrator, and Express and seek to establish new standards in AI image quality and user control.
The Firefly Image 2 Model, launched today at the Adobe Max conference and already available in beta on the Firefly web app, improves the quality and control of creative images while ensuring safe commercial use.
According to the official announcement, the updated version of the AI image generator now includes:
- Generative match is a feature that lets you set a visual style from pre-selected sample images or upload your own to generate new pictures that match the look and feel of your brand or project
- Photo settings are digital photography-style adjustments, similar to camera controls (such as depth of field or motion blur), that you can tweak for a more realistic and higher-quality result.
- Improved prompt capabilities can now autocomplete your prompts and recognize more landmark and cultural symbols.
- Prompt guidance gives our suggestions that will help you rewrite or add detail to your prompts to get the best possible image results.
- Share from Firefly function lets you share a link to your image creation, including its prompt, for others to see the process and even collaborate by reworking the results.
- Save to Libraries enables saving AI-generated images from Firefly directly to your Creative Cloud library so you can open it on any Adobe app of the ecosystem.
These improvements promise to let you create better and more realistic images than ever before. And equally important, it’s worth remembering that Firefly-created content is now fully cleared for commercial use, including for personal use and your brand or business.
Also, during Adobe Max 2023’s first day, Adobe disclosed that users have already created more than 3 billion images with the first version of this model since its launch in March.
The newly released Firefly Vector Model is a first-of-its-kind AI model that creates vector graphics from text prompts. It is the base for the new beta feature in Adobe Illustrator, the Text to Vector Graphic generator.
This tool generates “human quality,” fully editable and scalable vector graphics based on a text description, including icons, scenes, subjects, and patterns.
The AI-generated vectors come with advanced characteristics, such as editable gradients, seamlessly tile-able patterns, precise geometric lines, and grouped and organized layers to simplify posterior edits of the design.
Last but not least, the new Firefly Design Model generates impressive template designs from simple text prompts for various purposes.
Its functionality powers the newly added Text to Template function (in beta) in Adobe Express. It can create template designs for social, web, and print uses in all popular aspect ratios. The designs are fully customizable and automatically pick colors, text fonts, text sizes, images, and composition to achieve your design’s goals.
These advancements in the Firefly models aim to improve the quality and user control of creative content output in the Adobe suite.
The new models and features will benefit users, as they can now match existing image styles, enable more photorealistic image quality, better understand text prompts, create vector graphics and custom templates and share or save their work directly from Firefly.