Nano Banana Pro stole the show late at night, but the biggest highlight isn't AI image generation.

Ultraman is facing its darkest hour.

Google's AI offensive shows no sign of weakening. If a few days ago, Gemini 3 Pro extended its reach to the "front - end" field, today it's the turn of the design industry.

The newly released Nano Banana Pro (Gemini 3 Pro Image) has once again made a powerful impact in image - generation capabilities. It seems that junior designers' jobs are at stake.

The core features are as follows:

Resolution support: It can output images with a maximum resolution of 4K.

Multi - round editing: It supports a conversational, multi - round image - editing workflow.

Multi - image synthesis: It can combine up to 14 input images into 1 output image.

Search enhancement: It integrates Google's search capabilities, providing more accurate and up - to - date knowledge support.

No more "guessing blindly". Nano Banana Pro has finally learned to think before drawing.

Nano Banana's signature features are strong character consistency and a conversational editing method. The core evolution of Nano Banana Pro lies in its full integration of Gemini 3's in - depth thinking ability into the image - generation process.

Before generating an image, it will conduct a round of physical simulation and logical deduction, rather than just "guessing randomly" based on visual patterns.

Prompt: Please draw a four - panel picture. The four pictures should sequentially show the same young man wearing a straw hat pronouncing the words "I", "go up", "morning", "eight" respectively. The character's appearance should remain consistent, and the mouth shapes should accurately correspond to the pronunciation of each word. The overall style should be unified, with an aspect ratio of 16:9 and a resolution of 4K.

Cross - modal understanding is also more thoroughly demonstrated in Nano Banana Pro.

With the enhanced multi - language reasoning ability of Gemini 3, you can directly generate text in multiple languages, or localize and translate your content with one click.

A friend sends you a page of comics and asks the model to color the comics and translate the English in the speech bubbles into Chinese. Nano Banana Pro colors cleanly, with natural lighting and shadows. It accurately recognizes the text, and the English typesetting fits perfectly with the shape of the speech bubbles. The whole process, from recognition to translation and rearrangement, is carried out smoothly, as if it is truly "understanding" the picture.

Prompt: Translate the text on the picture into Chinese and color it, keeping other things unchanged.

Alternatively, for multi - language comics, international posters, and promotional materials that designers used to adjust repeatedly, they can now directly let AI handle it all at once. For example, let the model translate the English in an English poster into Chinese. This coherent processing method from recognition, translation to design is precisely the most powerful aspect of the native multi - modal architecture.

In terms of text - generation ability, through actual testing, although Nano Banana Pro may occasionally have inconsistent performance, its overall performance is excellent.

Prompt: Generate two vertical screenshots of a WeChat group chat interface, where group members are discussing the release of Nano Banana Pro.

I wonder if the cover deceived you? 🤯

Whether it's a short slogan or a whole paragraph of text, it can be clearly read, and it even supports fine - tuning of various textures, fonts, and calligraphy styles.

Prompt: In the style of ancient Chinese line - drawing illustrations, Guan Yu is sitting beside an oil lamp, wearing a wide - sleeved battle robe, with a focused and steady expression. On the table, there are bamboo slips of "Spring and Autumn", a gilded knife, a writing brush, and other items, outlined with delicate lines, retaining the ancient printing style. The background only sketches the corner of the wall, the screen, and the weapon rack with a few strokes, simple yet full of ancient elegance. The colors are mainly light ochre, gray ink, and light cyan, presenting the cultural charm and historical sense of ancient book illustrations, with an aspect ratio of 4:3.

The 64k input Token limit means it can understand extremely long text prompts. Whether it's a detailed storyboard script or complex multi - language typesetting requirements, it can understand them well.

Prompt: Generate a 4K ancient Chinese painting with the following poem written on it: "When will the bright moon appear? I raise my cup to ask the sky. I don't know what year it is in the celestial palace tonight. I want to ride the wind and return, but I'm afraid the jade palaces are too cold on high. I dance with my shadow, which seems more like being in the mortal world. The moon turns around the red pavilion, shines low through the embroidered windows, and shines on those who can't sleep. It shouldn't have any grudges. Why is it always full when people are separated? People may have sorrow or joy, be near or far apart. The moon may be dim or bright, wax or wane. This has been difficult since ancient times. I only hope that people can live long and share the beautiful moon thousands of miles apart."

To address the long - standing issue of low resolution in previous versions, Nano Banana Pro has upgraded the image quality to 4K and allows users to freely set more aspect ratios. Movie posters, widescreen wallpapers, and vertical storyboards can all be directly generated.

Nano Banana Pro also supports the combined editing of up to 14 input images while keeping the appearance of up to 5 characters consistent.

Combined with the multi - round dialogue ability, users can continuously adjust and integrate multiple materials until they achieve the desired effect. Whether it's turning a sketch into a product or converting a blueprint into a realistic 3D building, it can easily bridge the gap from concept to finished product.

Prompt: Doraemon and Li Bai are having a drink under the moon. The full moon is hanging high, and there are ancient pavilions and towers. Doraemon is wearing Tang - dynasty clothing, and Li Bai is holding a wine pot. There are wine utensils on the stone table, and it has a fairy - like atmosphere, with a mixed Chinese - Japanese painting style and delicate details.

What's more advanced is its professional - level creative control ability.

You can select, fine - tune, or transform any part of the image, from adjusting the camera angle, changing the style, to applying advanced color - correction, and even changing the scene lighting - turning day into night or creating a bokeh effect.

These tasks that used to require meticulous operations in Photoshop can now be achieved with just one sentence.

Search + Generation =? Google has given the ultimate answer.

If search is the "left brain" of Gemini 3, then image generation is its "right brain".

This is also an underestimated but most disruptive ability in the architecture of Nano Banana Pro (Gemini 3 Pro Image). Traditional search involves users searching, search engines providing links, users clicking into websites, and websites providing interfaces. Nano Banana Pro has introduced a search - enhancement feature (Grounding with Search).

When a user requests to generate a visual picture showing a 2 - day travel itinerary in Guangzhou, the picture generated by Nano Banana Pro includes a detailed itinerary map, Chinese and English annotations, and pictures of scenic spots.

For example, Nano Banana Pro can obtain the latest weather conditions from the search according to the prompt and convert key data such as temperature, wind force, humidity, and weather trends into vivid and design - oriented visual content.

Prompt: Search for real - time weather information in Guangzhou and create an infographic in the Chinese pop - art style, with an aspect ratio of 4:3.

This ability is important because it gives the creative process a factual basis, real - time nature, and verifiability. It can only be said that search is truly Google's forte, leading in both the depth of technological accumulation and understanding.

In terms of product positioning, Google has adopted a dual - model strategy: the old - version Nano Banana is for quick and fun daily editing, while Nano Banana Pro focuses on professional needs for complex compositions and top - level image quality. Users can choose freely according to the scenario.

For consumers and students, Nano Banana Pro is globally available in the Gemini app. Users only need to select "Generate Image" and enable the "Thinking" mode to use it. Free users will have a limited quota, and after exceeding it, they will automatically switch back to the original Nano Banana.