Apple prepares to add more AI models to Image Playground

Apple is preparing to expand Image Playground, its AI-powered image creation app, beyond its current integration with OpenAI’s ChatGPT model. According to code discovered in the latest iOS 26, macOS 26, and iPadOS 26 betas, Apple is laying the foundation to support multiple third-party AI models in the future.

When Image Playground launched, it relied on Apple’s own in-house models, offering basic styles such as “Birthday” and “Fantasy.” However, the output was considered limited compared to rival image generators. Apple later brought in ChatGPT’s styles to broaden creative options, allowing direct prompting of OpenAI’s model for more refined results.

Now, with new beta frameworks introducing “estimated latency” metrics and provider identifiers, Apple appears to be preparing the app to handle a wider range of external AI models. This could allow Apple to dynamically select the best-performing model based on latency or other factors.

While Apple has not confirmed which partners might be added, one strong candidate is Google’s Gemini 2.5 Flash Image, also known as Nano Banana, which has gained traction for its realistic image generation. Analysts also speculate that Apple may eventually allow users to integrate open-source models, though this would raise security concerns.