Google Whisk AI explained: How remixing works, availability, and how it differs from Gemini

Google has introduced a new AI experiment called Whisk, offering a unique approach to creating and visualising images using generative AI. Unlike traditional methods, which require submitting detailed, lengthy prompts, Whisk allows users to prompt with images instead. According to Google, all you need to do is drag and drop your images to start generating new ones. There are several nuances to how Whisk works, and here, we will explain its functionality, availability, and how you can remix images.

Google Whisk uses Gemini and Imagen 3 models behind the scenes

Google Whisk isn’t a brand-new AI model. Instead, it is just a tool that uses both Google Gemini and Google Imagen 3 to make images for you. But before that, let us explain how Whisk takes images as prompts. Firstly, you will need to enter an image for the subject, another image for the scene, and another image for the style. Then, what Whisk essentially does is remix the images, blending all three, and then creates an image that you could call your own. 

But behind the scenes, Google is actually using Gemini to write detailed prompts from the images you submit. Google Gemini, after analysing the images you submitted, writes detailed prompts and then submits them into Google’s Imagen 3 image generator.

Also Read: Google launches new AI image and video generation tools, Veo 2, Imagen 3, and Whisk- All details

Your images could differ slightly from reference material

Google knows that Whisk only captures your subject’s essence and not an exact replica—It only extracts a few characteristics from your image, and this is why the results may look different from what you expected. Laying out an example, Google says the generated subject might have a different height, weight, hairstyle, or even skin tone. Google says that it understands these characteristics may be important to what you are working on or generating, and this is why it lets you edit the prompts.

How is Whisk different from making images using Google Gemini?

Well, firstly, if you want to create images using Gemini, which also uses Imagen 3 behind the scenes, you have to submit a very long, detailed prompt to get a desired-looking image—and even then, it is not guaranteed that the AI will interpret it correctly or that your prompts will accurately describe what you are imagining.

Here, Whisk makes it easy to generate images because you are working with already created images. So, if you have a reference in mind, you can use those images to create an amalgamation or remix of sorts. This makes the image creation process a bit easier than writing traditional text-based prompts.

Also Read: Vivo X200 Pro vs Oppo Find X8 Pro: Which MediaTek Dimensity 9400 powered smartphone to buy?

Google Whisk: Availability

Unfortunately, Google Whisk is not available for those in India right now, or anywhere else for that matter, because Google Whisk is only available in the US for now. Users can already try it by visiting this link.