What Are Nano Banana Pro’s New Features? Full 2025 Breakdown

TL;DR

Nano Banana Pro improves Google’s AI image generation with stronger multilingual text rendering, more consistent product photography, sketch-to-photo editing, style transfer, and the ability to blend up to six uploaded images. Built on Gemini 3 Pro image, it is rolling out across Gemini, Workspace, Slides, Notebook LM, and Vertex AI, although access varies by region and subscription. Read on to see its capabilities, limitations, availability, and SynthID support.

Transcript

Google just dropped something that caught everyone offguard. Hot of the heels of Gemini 3, they released Nano Banana Pro, their most advanced image generation model yet. And I can finally show you what this thing is actually capable of. This isn't another incremental upgrade wrapped in the market in noise. This is the first time Google's image mode... Read More

Key Insights

• Nano Banana Pro is built on top of Gemini 3 Pro image, which is why it feels much stronger than the previous Nano Banana. It is presented as an evolution of Google DeepMind's top-tier image generation and editing model.
• Text rendering is significantly improved across multiple languages, though complex or long text may still produce occasional errors. Previous models constantly botched text with random letters, weird spacing, and gibberish, so this is a major leap forward.
• Nano Banana Pro can blend up to six uploaded images into a cohesive result, such as combining a color palette, typography sample, texture, product shot, lifestyle image, and logo into a unified brand identity board.
• The doodle-to-edit feature turns a rough sketch or annotation into a polished image, preserving composition exactly while adding realistic lighting, textures, and depth, which is useful for storyboarding and client presentations.
• Style transfer takes the visual style from one image and applies it to another, for example giving a basic headphones product shot a cinematic color grade, film grain, and dramatic lighting from a reference film still.
• Nano Banana Pro is rolling out across Gemini, Workspace, Slides, Notebook LM, and Vertex AI, with availability varying by region and subscription, and appears under the Gemini app's thinking toggle with limited pro generations on the free tier.
• A new Gemini feature lets you upload an image and ask whether it was generated by Google AI. SynthID identifies images made with Google models when the hidden marker is present, but it is not a universal AI image detector.
• Product photography is a standout use case. A luxury watch prompt tested 50 times with slight variations produced remarkably consistent professional results with accurate lighting, shadows, reflections, and marble texture.

Questions & Answers

Q: What is Nano Banana Pro, and what are its main new features?

Nano Banana Pro is Google DeepMind’s image generation and editing model built on Gemini 3 Pro image. Its highlighted features include improved multilingual text rendering, blending up to six images, converting rough sketches into polished images, style transfer, and more consistent product photography.

Q: How good is Nano Banana Pro at multilingual text rendering?

Its text rendering is significantly improved across multiple languages, although complex or lengthy text can still contain occasional errors. Tests included logos using different languages and a bilingual English-and-Japanese café menu with ingredient lists, prices, and clay-style icons.

Q: How many uploaded images can Nano Banana Pro blend?

Nano Banana Pro can combine up to six uploaded images into one cohesive composition. One test blended a color palette, typography sample, texture, product shot, lifestyle image, and logo concept into a unified brand identity board.

Q: How does Nano Banana Pro perform for product photography?

Product photography was a standout use case in the testing. A luxury mechanical watch prompt produced professional-looking lighting, shadows, reflections, and marble texture, while 50 tests with slight variations showed strong consistency; an earbuds test also produced clean, e-commerce-ready results.

Q: Can Nano Banana Pro turn a rough sketch into a polished image?

Yes, its doodle-to-edit capability can transform a rough sketch or annotation into a polished image. It can preserve the original composition while adding realistic lighting, textures, and depth, making the feature useful for storyboards and client presentations.

Q: How does style transfer work in Nano Banana Pro?

Style transfer applies the visual treatment of a reference image to another image. For example, a basic headphones product shot can adopt cinematic color grading, film grain, dramatic lighting, and other qualities from a reference film still.

Q: Where is Nano Banana Pro available?

Nano Banana Pro is rolling out across Gemini, Workspace, Slides, Notebook LM, and Vertex AI, with availability depending on region and subscription. In the Gemini app, it appears under the thinking toggle, and the free tier includes limited Pro generations.

Q: What can SynthID detect in Gemini?

Gemini lets users upload an image and ask whether it was generated by Google AI. SynthID can identify images made with Google models when its hidden marker is present, but it is not a universal AI-image detector; Google AI Ultra users can also receive images without the visible Gemini watermark while an invisible watermark remains.

Summary & Key Takeaways

• Google DeepMind launched Nano Banana Pro, its newest and most advanced AI image generation and editing model, built on top of Gemini 3 Pro image. It takes existing AI image use cases like mock-ups, posters, logos, and product shots and makes them cleaner, more controllable, and easier to prompt.

• The model significantly improves multilingual text rendering, blends up to six uploaded images into cohesive compositions, converts rough sketches into polished photographs, and applies style transfer from a reference image. Product photography tests for watches and earbuds produced consistent, commercial-grade results ready for e-commerce listings.

• Nano Banana Pro is rolling out across Gemini, Workspace, Slides, Notebook LM, and Vertex AI, with availability varying by region and subscription. A SynthID-based tool can detect Google-generated images, and the visible Gemini watermark is being removed for Google AI Ultra users while keeping an invisible watermark.