Gemini 3 Pro And Nano Banana 2 Could Launch Together

By Chris Smith
The Gemini app logo on an Android phone. Proxima Studio/Shutterstock

OpenAI on Wednesday surprised users with the ChatGPT 5.1 release. The update was expected, as it appeared in some leaks ahead of its official rollout, but it still seemed like a sudden decision from OpenAI. Fans of generative AI software are probably aware of rumors that say Google might soon unveil the Gemini 3.0 (also styled Gemini 3) upgrade and the new Nano Banana 2 image generation model. Both were spotted in testing over previous weeks, with Nano Banana 2 going viral over the weekend. Rumors said Google might unveil the Nano Banana 2 as soon as this week, which would explain OpenAI's rush to make ChatGPT 5.1 available to users.

While that's all speculation, it turns out that Google might have planned a big launch of new AI models. Code found in the Gemini iPhone app has the following string of text in it, as spotted by X users who keep track of new software features, including AI tools: "Try 3 Pro to create images with the newer version of Nano Banana." This appears to be a marketing message that will be available in the Gemini app, instructing users that they can use Gemini 3.0 Pro with the new Nano Banana 2 model. Nothing is official until Google announces it, but if you've been following the recent leaks, you might have realized why the Gemini 3.0 models need the new Nano Banana 2 upgrade.

Coding apps with Gemini 3 and Nano Banana 2

Using Nano Banana inside the Gemini web app. Nwz/Shutterstock

A few weeks ago, someone posted videos online that show Gemini 3.0 can clone operating systems including iOS and macOS (below). That is, you can instruct the AI to make a functional clone of smartphone and computer operating systems, and Gemini will do it, in the version people say they tested online.

Such built-in Gemini 3 capabilities could open the door to new types of vibe-coding where anyone can create apps without worrying about designing the user interface. If the image generation abilities seen in last month's leaked demos are accurate, Gemini will be able to create high-quality visuals for app user interfaces by following closely the user's instructions. A detailed text prompt might be all Gemini needs to create an app interface with menus, buttons, and media the user can interact with while testing the app.

For that to happen, Gemini 3 will need access to powerful AI image and video generation models. In October, we'd have assumed Nano Banana and Veo would allow Gemini to clone an existing operating system with such fidelity. But now that Nano Banana 2 samples have gone viral, including images where the AI clones the Windows operating system (below), it makes sense to let the upgraded Nano Banana model power any Gemini 3 tasks that involve creating app experiences featuring polished visuals.

That said, we're only speculating based on what we have seen in these recent discoveries. We'll need to wait for Google to confirm everything. That might happen sooner than later, given the surprise ChatGPT 5.1 rollout.

Recommended