Gemini 3 is back to make a splash! Google's popular live image model upgrade, Nano Banana Pro offers "studio-level" precision and control

Google's most powerful artificial intelligence (AI) model, Gemini 3, just two days after its release, is back with a surprise.

On Thursday, the 20th, Eastern Time, Google announced an upgrade to its popular image generation and editing model based on Gemini. The newly launched model, Nano Banana Pro, claims to provide "studio-level" precision and control, achieving significant improvements in text rendering accuracy, resolution, and professional control capabilities, aiming to push AI image tools from consumer applications into the professional design field.

Nano Banana Pro supports generating images with a maximum resolution of 4K, can accurately render text in a multilingual environment, and allows users to control parameters such as shooting angle, depth of field, color, and lighting like using a professional camera. Josh Woodward, Vice President of Google Labs and Gemini, stated that the product excels in infographic creation, slide design, and can maintain consistency for up to five characters and fourteen objects.

Google's global free Gemini users will be able to use Nano Banana Pro within quota limits, and once the quota is exceeded, it will automatically switch back to the old model. Paid AI plan subscribers will have higher usage limits. The model has been integrated into mainstream design tools such as Canva, Figma, Adobe Firefly, and Photoshop.

Google also launched an AI content traceability feature simultaneously. Users can upload images in the Gemini app to check whether they were generated by Google AI, and this feature will expand to audio and video in the future. This move is Google's latest attempt to commercialize AI technology and an important step in catching up with OpenAI in the generative AI race.

This release continues Google's strong momentum in the AI field this week. The Gemini 3 released on Tuesday achieved a "huge leap" in reasoning and coding capabilities, pushing the stock price of Google's parent company, Alphabet, to a historic high on Wednesday. On the day of the official announcement of Nano Banana Pro, the stock price rose nearly 5% in early trading but followed the market down in the afternoon, closing down about 1%, temporarily parting ways with the record high closing price set on Wednesday.

Tackling the AI Spelling Challenge, Breakthrough in Text Rendering

A core challenge that AI image generation models have long faced is text rendering accuracy. Although some models have improved over time, spelling errors and font distortions remain common when generating images containing multiple text phrases, limiting their potential as professional design tools.

Nano Banana Pro directly addresses this issue. A Google spokesperson stated that the new model can better plan the text position, font characteristics, and their spatial relationships with other image elements before rendering the final image This technology can convert recipe text into illustrated flowcharts or visualize real-time information such as weather and sports.

Nano Banana Pro supports generating clear and distinguishable text, allowing users to describe the desired font type or simulate different handwriting styles, and also generate localized text or translated words within images. This enables users to create product displays, posters, and infographics targeted at international markets. Google emphasized in a blog post that sharp, clear text helps users create impactful posters, complex charts, and detailed product models.

However, Google also reminds users to be aware of limitations. The new model may misinterpret information or produce factual errors when generating infographics, labeling charts, or presenting complex data. While it can generate and translate multilingual text, issues may still arise in grammar, spelling, cultural nuances, or idioms. Users should always carefully check the accuracy of the generated images and their accompanying text.

Professional-level control features, resolution upgraded to 4K

Nano Banana Pro offers unprecedented fine control capabilities for professional users. Users can explore different shooting angles and types, including wide-angle, panoramic, and close-up, and can adjust depth of field to focus on different subjects within the image.

In terms of color and lighting, users can adjust color grading and lighting direction, and even switch scenes from day to night.

The model supports precise scaling in 1K, 2K, and 4K resolutions, and can switch aspect ratios freely to fit different platforms or purposes.

The cost of increased resolution comes with higher costs and slower speeds. The original Nano Banana model generated 1024-pixel images at a cost of $0.039, while the new model generates 1080p or 2K images at a cost of $0.139, and 4K images at $0.24.

In terms of material consistency, the model can maintain consistency and similarity for up to five characters and fidelity for up to fourteen objects within a single workflow. For brands looking to incorporate their own designs into marketing campaigns, the model can accept up to 14 reference images and rearrange them in new scenes described by user text prompts while retaining the features of the input materials The model can also transform sketches into objects, convert hand-drawn notes into charts, and turn ideas into 3D rendered buildings. Users can create multiple images at once, quickly and efficiently exploring and reviewing creative options. Google stated that the model can also create realistic images of landscapes, plants, people, and animals with true details.

Extensive Product Integration, Paid Users Get Higher Quotas

Nano Banana Pro is being fully promoted within Google's existing AI tools. The Gemini App will default to using the new model to generate images, with free subscription users able to use it within a limited quota; exceeding this will switch back to the original Nano Banana model. Google AI Plus, Pro, and Ultra subscription users will receive higher generation limits, and these paid users can also use the model in Notebook LM.

In the United States, AI Pro and Ultra subscription users can access the model in search through AI mode. Ultra subscription users can also use it in Google's video tool Flow, while Workspace customers can use it in Google Slides and Vids. Developers can call Nano Banana Pro through the Gemini API, Google AI Studio, and the new IDE Antigravity.

Google is also integrating its SynthID technology into the Gemini application to add watermarks and detection for AI-generated images. Users can upload images, and the chatbot will inform them whether the image was created or modified by Google's image model. Google currently creates invisible digital watermarks for all media generated by AI tools, and images created by free or Pro plan subscribers will also have visible watermarks, while users subscribed to the highest-priced Ultra plan can remove the visible watermarks. Google did not mention whether it plans to support other AI watermark standards like C2PA.

Strong User Growth, Accelerated Commercialization Process

The predecessor of Nano Banana, Nano Banana, quickly became popular on social media after its launch in August this year, allowing users to transform photos of themselves or their pets into hyper-realistic 3D mini sculptures. Woodward wrote in a post on X in September that the product helped the Gemini app gain 13 million new users in just four days.

According to data released by Google, the Gemini App currently has over 650 million monthly active users, with AI Overviews powered by Gemini reaching 2 billion monthly users. In comparison, OpenAI CEO Sam Altman stated in October that ChatGPT has 800 million weekly active users.

Woodward stated this Thursday that the demand for Google's AI products continues to grow, with many users signing up for Gemini subscription plans to gain "higher limits for these advanced models." He noted, "We are seeing a large influx of users towards these products, which is actually the best problem to have—there is high demand, and we are working hard to figure out how to serve that demand." This release marks Google's latest attempt to commercialize AI technology. Since the launch of ChatGPT in 2022, which sparked a generative AI race, Google has been striving to catch up with OpenAI. Last week, OpenAI announced two updates to the GPT-5 model, making it 'default warmer and more conversational' and 'more efficient and easier to understand in everyday use.' Currently, ChatGPT ranks first on the Apple App Store's free app chart, with Gemini in second place.

Woodward stated that the company plans to continue expanding its AI product offerings, focusing on AI filmmaking tool Flow and the 'world-building' model Genie, which is currently available as a limited research preview