• Home
  • digital
  • Google’s ‘Nano Banana 2’ images go viral ahead of launch, sparking buzz online

Google’s ‘Nano Banana 2’ images go viral ahead of launch, sparking buzz online

The new model will reportedly demonstrate major improvements in text rendering, infographics, charts, world knowledge, and adherence to user instructions.

By  Storyboard18| Nov 11, 2025 1:00 PM
Follow us
Google’s ‘Nano Banana 2’ images go viral ahead of launch, sparking buzz online
The new Nano Banana model will reportedly demonstrate major improvements in text rendering, infographics, charts, world knowledge, and adherence to user instructions.

Google’s much-anticipated image-generation model, Nano Banana 2, has already become a viral sensation — even before its official launch. The follow-up to Google’s Gemini 2.5 Flash model, known popularly as Nano Banana, is drawing widespread attention across social media for what users claim are leaked examples of its vastly improved image-generation capabilities.

The original Nano Banana, launched a few months ago, impressed users with its ability to perform complex image edits and generate visuals from natural-language commands. However, the latest round of online chatter suggests that Google’s next iteration is set to go further, with substantial upgrades in image accuracy, rendering quality and instruction following.

According to a post by Testing Catalog News on X (formerly Twitter), the new model will reportedly demonstrate major improvements in text rendering, infographics, charts, world knowledge, and adherence to user instructions. Some users even claim that the Nano Banana 2 briefly appeared on select online platforms before being swiftly taken down by Google.

Early leaks indicate that the model could be launched as early as this week, and may feature broader support for high-resolution outputs such as 1K, 2K, and 4K, along with multiple aspect ratios, including 9:16 and 16:9. This would address one of the most frequent criticisms of the first Nano Banana — its difficulty in maintaining correct aspect ratios. Competing models like Qwen currently allow users to select resolutions during generation, something Google’s model lacked until now.

Adding to the intrigue, industry rumours suggest that Nano Banana 2 could be powered by Google’s upcoming Gemini 3.0 architecture, marking a significant leap in AI imaging performance and creative control.

Although Google has not officially confirmed the model’s existence or release timeline, social media platforms are already flooded with images allegedly generated by Nano Banana 2. These examples appear to showcase remarkable progress in accuracy and contextual understanding.

One common flaw among most AI image generators has been their inability to depict accurate time on analogue clocks, with nearly all outputs defaulting to the familiar 10:10 display used in commercial photography. Users claim Nano Banana 2 has overcome this limitation, now producing clocks showing varied and realistic times.

Others have shared examples of the model’s UI and webpage generation capabilities, where the AI is said to produce nearly flawless webpage mock-ups and browser screenshots directly from textual prompts — a task that usually results in distorted layouts or garbled text in other systems.

Viral posts also show images of world leaders including Donald Trump, Volodymyr Zelenskyy, and Vladimir Putin, allegedly generated by the model on an old-style monitor screen. Another example circulating online shows Nano Banana 2 creating a classroom scene, where the AI-generated teacher solves problems and writes answers on a whiteboard with impressive visual coherence.

While there is no word yet from Google on whether Nano Banana 2 was unintentionally released or deliberately tested on limited platforms, the online excitement and speculation underscore the growing anticipation surrounding Google’s next major step in generative AI imagery.


Tags
First Published on Nov 11, 2025 1:08 PM

More from Storyboard18