Google expected to launch Nano Banana 2 soon: Here’s what we know about it

Gemini

Google’s quirky yet powerful AI image generator is getting an upgrade. The company is reportedly preparing to launch Nano Banana 2, the next version of its experimental image model that’s part of the Gemini app. A preview of the tool briefly appeared online before being shared widely on X (formerly Twitter), giving users a sneak peek at what’s coming — and it looks like Google is taking a serious step forward in how AI creates visuals.

So, what exactly is Nano Banana 2? At first glance, it might sound like a lighthearted name, but behind it lies a sophisticated AI model that aims to make image generation more human-like and intelligent. According to early reports, Nano Banana 2 will offer better control over camera angles, viewpoints, and image precision, especially when it comes to colors and lighting. It will also let users edit text inside images — a feature many creators have long wanted — without disturbing the rest of the picture.

But the most interesting part is how it thinks. The new model appears to adopt a multi-step workflow, meaning it doesn’t just generate an image in one go. Instead, it plans the image first, analyzes it for mistakes, and then refines it through self-correction until the final version is ready. In simple terms, it’s as if the AI is learning to be its own art director — sketching drafts, catching flaws, and polishing details before showing you the finished result.

That shift marks a big philosophical change for Google’s AI tools. Rather than being reactive, Nano Banana 2 is designed to be proactive — noticing and fixing its own errors before users do. This mirrors how human artists and designers work, and it could make AI-generated visuals feel far more natural and polished.

Leaked previews show cleaner lines, better perspective, and fewer of the awkward distortions that have long given away AI art. Some code references also mention something called Nano Banana Pro, hinting that Google may already be working on a premium version for higher-resolution or commercial use.

Interestingly, the new model — also known internally as GEMPIX 2 — has started showing up in other experimental Google projects like Whisk Labs, which focuses on creative AI integration. If it follows the same rollout pattern as the first Nano Banana, users may soon notice an upgrade in how their photos and illustrations appear across Google’s ecosystem.

In short, Nano Banana 2 isn’t just about sharper images — it’s about smarter ones. By blending creativity with self-awareness, Google seems ready to make AI-generated art feel less robotic and more like something a human might make.