With macOS Tahoe 26, iPadOS 26, and iOS 26, Apple expanded its partnership with OpenAI to include ChatGPT as a generative image provider on Image Playground. Now it is setting the stage to bring in even more third-party models.
When Apple debuted Image Playground, the app leveraged the company’s own models to let users create illustrations based on prompts, or on their contacts.
The app offered predefined styles and themes, such as “Birthday,” “Fantasy,” or “Starry Night. “ However, compared with other image generation models, the end result looked rather primitive.
For this reason, on iOS 26, macOS 26, and iPadOS 26, Apple expanded its partnership with OpenAI, and brought in ChatGPT styles, which offer their own preset filters to get users started. There is also the option to directly prompt OpenAI’s model, in case none of the presets deliver what the user has in mind.
Image Playgrounds to offer an even wider selection of third-party models
Based on code introduced in today’s round of betas, 9to5Mac can confirm that Apple is laying the groundwork for an expansion of third-party models on Image Playground, beyond OpenAI’s model.
With today’s betas, Apple added a new metric called “estimated latency,” as well as provider identifiers, to the Image Playground framework that previously handled only the OpenAI model.
This could mean that Apple may choose the best model for the task based on the latency of each one at any given time, or it may mean that Apple wants to have this metric handy for its internal telemetry.
But more importantly, it means that Apple is definitely changing the underlying structure of the app to support multiple third-party models beyond OpenAI’s, which is great news for users of the app.
When can we expect new models?
There is no word yet on which models may make their way into Image Playground, but one strong candidate could be Google’s Gemini 2.5 Flash Image, also known as Nano Banana.
The company’s new model has gone viral in recent weeks for its hyper-realism and its ability to convincingly recreate people’s likenesses from user-submitted images. This led the Google Gemini app to the top of the App Store in multiple countries, including the US.
It could also be the case that Apple may let users download open-source models and plug them into Image Playground (much like on Xcode), although that could be a stretch, given the potential for misuse.
Apple goes through great lengths to ensure Image Playgrounds works as a safe image creation tool, so it is more likely that the changes introduced in today’s betas point to an expansion of Apple’s partnerships.
Do you have a favorite AI image generator? Let us know in the comments.
Accessory deals on Amazon