Google launches an aificial intelligence model that produces professional images at high speed

The new system, whose official name is “Gemini 3.1 Flash Image,” combines two basic features, as it combines the high speed that was provided by the “Gemini Flash” model with the accuracy and quality of information that was available in the previous version, “Nano Banana Pro.” This combination aims to produce high-quality images without sacrificing speed.

The new model is distinguished by its ability to better understand instructions, as it can execute complex commands with high accuracy and understand the precise details requested by the user. It also provides accuracy in drawing text, as it can write clear texts within images, which makes it ideal for creating greeting cards, marketing ads, and various design templates.

Another striking feature is that it can translate texts in images into other languages ​​while preserving their original form, which facilitates its use in international projects.

The new model allows users to have full control over the quality of the images, as they can choose the image resolution from 512 pixels up to high 4K quality, which is suitable for social media posts or large display screens.

It also keeps different elements consistent, as it can preserve the features of five different characters or 14 elements in multiple images in one workflow.

The original “Nano Banana” version appeared last year, and was famous for its ability to edit and create images, then came the more advanced “Pro” version with higher resolution. Now with the “Nano Banana 2”, it offers a combination of the speed of the first version and the accuracy of the second version in one faster and more efficient model.

The model is available across several Google platforms, where it can be used within the Gemini application in its various modes, through Google Search (AI Mode) and Lens, and in AI Studio and the Gemini API. It is also available on the Vertex AI platform in Google Cloud, and in the Flow application for creating images for free, in addition to its use in Google ads.

In a related context, Google continues to develop systems for recognizing content produced with artificial intelligence, as the company uses SynthID watermarking technology with the C2PA system to document the source of images. Since this feature was launched last November, it has been used more than 20 million times to identify artificial intelligence content.

By launching Nano Banana 2, Google is offering the technology of creating fast and accurate images as a basic feature available to everyone, and not as an additional paid service.

Source: interestingengineering