Google Whisk is a cutting-edge generative AI tool that enables creativity through image-based prompts. With Whisk, users can combine multiple images for subjects, scenes, and styles, remixing them into unique designs like stickers, plushies, or pins. Powered by the Gemini model and Imagen 3, Whisk is built for rapid creative exploration rather than precise edits, making it an exciting tool for artists and creatives.

Google Whisk offers a fresh approach to image generation by shifting focus from text prompts to image-based inputs. Simply drag and drop images to set your subject, scene, and style, then blend them to create personalized outputs. Whether designing an enamel pin or a digital plushie, Whisk empowers users to explore endless creative possibilities.

Behind the scenes, the Gemini model analyzes and captions your input images. These captions are processed by Google’s Imagen 3 model, which generates unique visuals inspired by the key characteristics of the uploaded images rather than replicating them exactly. This flexibility enables users to remix their inputs into novel combinations effortlessly.

While Whisk excels in creative exploration, it may not always align perfectly with user expectations, such as altering a subject's hairstyle or proportions. To address this, Whisk allows users to view and edit the underlying prompts, providing more control over the output.