Google is set to launch Nano Banana 2, building on the positive feedback for the original model in the Gemini app. The new version significantly improves capabilities in areas like precise coloring, view control, and text correction in generated images, addressing the shortcomings of Nano Banana 1. Leaked images show a noticeable enhancement in visual quality. Interestingly, Nano Banana 2 appeared on the Media AI platform in a preview, with outputs consistent with earlier short-term tests on Gemini. Internal sources suggest that Nano Banana 2 is still using the Gemini 2.5 Flash framework instead of transitioning to Gemini 3.0 Pro. This choice might allow for an earlier release while potentially upgrading to the newer model later. The model introduces a multi-step generation process that involves planning, generating, reviewing, and correcting images, enhancing accuracy significantly—claims indicate up to three times better performance than earlier iterations. The recent GitHub activity hints at a rebranding to “Nano Banana Pro,” reflecting these advancements. Feature-wise, Nano Banana 2 will support a variety of aspect ratios, including 1:1, 2:3, and 16:9, and output resolutions like 1K, 2K, and 4K. There's ongoing speculation about whether it will be built on Imagen 4 or continue with Gemini 3 Pro or Flash. The community is also discussing multiple product variants, including a high-resolution Nano Banana Pro. Official announcements are anticipated soon, as internal testing is in progress and cards have appeared in the Gemini UI, indicating a public rollout could happen within days. Google plans to extend the model's integration beyond the Gemini app to other platforms, which could enhance the capabilities for users across various creative and professional domains.