Google DeepMind has just introduced a major step forward in the field of generative artificial intelligence (GenAI): the Gemini 2.5 Flash Image model.

Affectionately known as “Nano Banana” by the community, this model is a powerful tool, capable of transforming text into images and editing existing images with great precision and flexibility.

This is a continuation of the Gemini model line, but is specifically optimized for image-related tasks.

c732a0fbdcce57900edf.jpg
The keyword Gemini topped the trending list on Google Trends in Vietnam on September 4. Screenshot.

With outstanding advantages in photo creation and editing capabilities, Nano Banana is attracting great attention from global technology users, including Vietnam.

According to Google Trends, searches for Gemini and Nano Banana have skyrocketed in the past 24 hours. Statistics from September 4 show that the keyword “Gemini” topped the Trending chart across all topics.

On forums and social networking groups, users also share and test image creation commands and scripts using Nano Banana.

Most of them highly appreciate the ability to create images "fast, neat, and agile" as well as the quality of the output product of the tool.

More than just creating images from text descriptions, Gemini 2.5 Flash Image excels in several core aspects: on-demand image editing, character consistency, impressive processing speed, and SynthID security technology.

Users can use natural language to change elements in a photo. For example, just type in the command “add a red chair next to the table” or “change the background to a forest” and the model will make the edit accurately and seamlessly.

Character consistency is considered the most groundbreaking feature. The model can memorize and recreate an image of a character or object with consistent features (like face, clothes, style) in many different photos.

This is especially useful for creative projects that require creating a series of related images, from comics to marketing campaigns.

Original photo and photo after editing with Gemini 2.5 Flash Image

With its optimized architecture, Gemini 2.5 Flash Image can produce high-quality results in seconds, significantly accelerating the creative workflow.

All images created or edited by the model are tagged with a proprietary invisible digital watermark from Google, allowing the recognition of AI-generated images, contributing to ensuring transparency and origin of the content.

The Gemini 2.5 Flash Image model is opening up a wide range of application potential, from individual users to professional developers.

Developers can access the model through the Gemini API and Google AI Studio to build their own applications, while businesses can use the model on the Vertex AI platform.

For the average user, the model is built directly into Google Gemini and other apps, making AI imaging an accessible and intuitive tool.

With its outstanding features, especially the ability to maintain character consistency, Gemini 2.5 Flash Image is expected to become a formidable competitor to traditional photo editing software such as Photoshop, changing the way creators and users interact with digital images.

China unexpectedly surpasses the US at the global AI translation competition . Tencent's open-source Hunyuan-MT-7B model (China) won almost completely over the US 'giants' at the international machine translation competition within the framework of the WMT25 conference.

Source: https://vietnamnet.vn/nano-banana-la-gi-ma-khien-moi-nguoi-xon-xao-dung-dau-google-trends-viet-nam-2439283.html