Adobe hasn’t shared a particular date but, however Firefly Picture Mannequin 5 will launch “within the months to come back” following Adobe MAX. Firefly Picture Mannequin 4 launched in April 2025, so it is going to possible be 2026 earlier than we see a broad launch of Picture Mannequin 5. Like Picture Mannequin 4, the up to date mannequin has a local 4-megapixel decision, which means it could actually generate photos at 2K (2560 x 1440). It additionally helps prompt-based enhancing, which sees edits generated at 2 MP or Full HD (1920 x 1080). The massive enchancment for Picture Mannequin 5, nevertheless, is layered picture enhancing.
In a demo, Adobe confirmed me how the brand new Firefly mannequin works. You’ll be able to add a picture, and Firefly Picture Mannequin 5 will determine completely different components, permitting you to maneuver, resize, and exchange these components with generative options. In my demo, Adobe used a bowl of ramen, exhibiting how Picture Mannequin 5 was capable of minimize out and transfer the chopsticks to a distinct space of the scene, in addition to add a bowl of chili flakes generated by AI. And all of this was executed with none visible artifacts.
New Generative Options in Adobe Firefly
Firefly can also be getting two new generative AI options: Generate Soundtrack and Generate Speech. Each do precisely as their names counsel, with some particular guardrails in place.
Generate Soundtrack will scan an uploaded video and counsel a immediate for a soundtrack. Moderately than deleting the immediate and ranging from scratch, Adobe enables you to select a vibe, fashion, and function to search out one thing that works. As an illustration, you would say you need a tense orchestral rating to put excessive of a chase scene.
Courtesy of Adobe