A new image model called Nano Banana Pro (Gemini 3 Pro Image) is now available for developers who want higher-quality image generation. It builds on Gemini 3 Pro and comes a few months after the release of Nano Banana (Gemini 2.5 Flash Image). Since that earlier release, the developer community has used its features in many ways, including keeping characters consistent, restoring photos, and making focused edits in large canvases.
The new model is rolling out in a paid preview and is designed to support multimodal apps through the Gemini API, Google AI Studio, and Vertex AI. It produces sharper images, handles text in images with better accuracy, and draws on broad knowledge. When grounding with Google Search is turned on, it can also pull in web content that matches the user’s prompt.
Gemini 3 Pro Image is also being introduced in more developer tools. In Google Antigravity, the company’s agent-driven development platform, coding agents can now use this model to create detailed UI mockups or new visual assets before any code is written. Creative platforms like Adobe and Figma are also adding support.
High fidelity and more control
For teams building tools that need precision, Gemini 3 Pro Image offers control over lighting, camera settings, focus, colour, and layout. The controls help developers produce visuals that meet professional standards.
The model supports 2K and 4K output, making it suitable for production use. It can combine many elements into one cohesive design, like product photos, logos, and references. It can maintain consistent appearance for up to five people, merge six high-fidelity inputs, or blend up to fourteen standard images into a single finished piece. A demo app shows how logos and product images can be paired to create mockup designs.
Clearer text and easier localisation
The model delivers a clear improvement from 2.5 Flash Image when working with text. It handles language and logic more reliably and produces clean, readable text inside images. This makes it useful for marketing content, educational material, and other projects that rely on accurate text.
The capabilities can be seen in the comic book generator app in Google AI Studio, where users can create multi-page comics featuring themselves and a friend, complete with styled text and layouts.
The model also supports more natural localisation. It understands the meaning of elements in an image, so language on signs, menus, or documents can be changed while keeping the original style and layout through image-to-image generation.
Gemini 3 Pro Image draws from a broad base of information, helping it produce visuals that are more accurate than those from earlier models. With Google Search grounding enabled, it can use real-time web data for more factual output. This is helpful for visuals that require precision, like diagrams or maps. A demo app lets users create infographics on any topic with content tailored to their needs.
Getting started with the model
The release includes many ideas and suggestions gathered from developers. Every generated image now includes a SynthID digital watermark, which identifies when AI was used.
Developers can start by exploring the collection of demo apps that use Gemini 3 Pro Image. They can then adapt those apps or bring the model into their own projects through the Gemini API in Google AI Studio or Vertex AI. Documentation, a prompt guide, a cookbook, and the developer forum are available for technical support throughout the process.
See also: Gemini 3: Google enables new agentic AI workflows for developers
Want to dive deeper into the tools and frameworks shaping modern development? Check out the AI & Big Data Expo, taking place in Amsterdam, California, and London. Explore cutting-edge sessions on machine learning, data pipelines, and next-gen AI applications. The event is part of TechEx and co-located with other leading technology events. Click here for more information.
DeveloperTech News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.