Google unveils Pixel Studio, a new app for AI image generation

While the Pixel 9 smartphone lineup was the highlight of today’s Made by Google event, the company also announced a new app to create AI images. Dubbed Pixel Studio, the app uses Google’s Imagen 3 model to transform text-based prompts into images. Beyond basic image creation, the app offers on-device AI editing tools that allow you to reposition or erase objects within the frame. It will also suggest art styles, such as cinematic and anime, depending on the prompt.

Many of the best AI image generators are locked behind an expensive monthly subscription. However, with Pixel Studio, you won’t have to pay for the feature beyond the initial cost of owning a Pixel 9 series device. It’s unclear if the app will make its way to older Pixel devices.

Pixel Studio uses Google’s own Imagen 3 model under the hood. Like Midjourney and other AI image generators, Imagen 3 relies on a machine learning technology called diffusion that allows it to convert text-based prompts into images.

While an on-device image generator sounds impressive, it’s worth noting that even Google’s latest Tensor G4 SoC cannot match the processing power of dedicated AI hardware running in the cloud. This is why the actual image generation still requires Imagen 3 and only a smaller on-device model handles edits after teh fact, similar to how Gemini Nano is cut down from its larger counterparts.

You can still create AI-generated images on older Pixel devices or other Android devices via the Gemini overlay on mobile. Simply ask the chatbot to generate an image. This functionality has been available for a while now, but it requires an internet connection and won’t produce results as quickly as the new Pixel Studio app.

The chatbot interface also doesn’t offer any editing features once a set of images has been generated. However, Google also announced improvements to the Gemini overlay today, allowing you to drag and drop AI-generated images directly into an email or chat.

The Pixel Studio app comes just weeks after Apple announced its own AI image generator for the iPhone. Dubbed Image Playground, the app is expected to arrive alongside other Apple Intelligence features that compete with other aspects of Google Gemini. Apple’s Image Playground will use an on-device model to generate images offline.