Nano Banana 2.1 is now available in Picsart

Google released Nano Banana 2.1 in October 2026, and it is now available in Picsart AI Playground. The new Nano Banana model generates and edits images from text and image input, outputs in 1K, 2K or 4K across 15 aspect ratios, and supports multi-turn editing, so a picture can be refined over a whole conversation instead of rebuilt from a fresh prompt every time. In Google’s own side-by-side human evaluations, it scored ahead of both Nano Banana 2 and Nano Banana Pro in every text-to-image and editing category reported.

What is Nano Banana 2.1?

Nano Banana 2.1 is the latest image model in Google’s Nano Banana line and a member of the Gemini 3 series of natively multimodal reasoning models. It is built on Gemini 3.6 Flash. It takes text and images as input, and it can return image and text output in the same response. Google aims it at four kinds of work: creating and editing images with professional precision and control over multiple quick iterations, generating clear text for posters and intricate diagrams, drawing on real-world knowledge, and rendering localized text across several languages.

Generate and edit in one conversation

Editing carries as much weight as generation in this release. Nano Banana 2.1 supports multi-turn image editing, and Google evaluated it on both static and dynamic conversations. A session can run as a back-and-forth: generate a first version, ask for a change, look at the result, and ask for the next one.

Edit with a doodle

Ink-based editing is one of the editing modes Google built and tested the model for. Draw on the picture to direct the edit. Mask and ink-based editing is also one of the categories where the new model pulled clearly ahead of Nano Banana 2 in Google’s tests. Google notes that instruction following and ink persistence in doodle edits are still partial, so a tricky edit can take more than one pass.

Objects, environments, styles and characters

Google’s editing evaluations cover general image editing, stylization, character editing, and object and environment editing. Swap an object, move a scene into a new environment, restyle the whole image, or change something about a character. Character consistency between the input images and the output is not always perfect, and Google reports rare cases where an edit keeps the pose of the original subject.

Several images, one result

Multi-image editing covers multi-product recontextualization and multi-character scenes. Bring several product shots into one new scene, or place several characters in the same frame. Multi-character consistency is also where the new model posted its largest score gain over Nano Banana 2.

Resolutions and formats

Output comes in 1K, 2K or 4K resolution, across 15 aspect ratios that run from square to wide banner shapes.

The 15 aspect ratios

• Square and near-square: 1:1, 4:5, 5:4, 3:4, 4:3
• Classic photo: 3:2, 2:3
• Widescreen and vertical: 16:9, 9:16
• Cinematic: 21:9, 9:21
• Banner: 4:1, 1:4, 8:1, 1:8

Text inside the image renders in several languages, and the model draws on Gemini’s real-world knowledge, which Google points at work like complex infographics and historically accurate scenes. Small text is the weak spot Google calls out: it is often blurry at 1K, and long paragraphs and full pages of text remain hard.

What changed since Nano Banana 2

This Nano Banana update shows up most clearly in Google’s scores. Google ran side-by-side human evaluations, rated as Elo, against Nano Banana 2 and Nano Banana Pro. With thinking on, Nano Banana 2.1 came out ahead of both in every reported category. A few of the results:

• Overall text-to-image preference: 1050, against 990 for Nano Banana 2 and 935 for Nano Banana Pro
• Multi-character consistency: 1106, against 978 and 1011
• Mask and ink-based editing: 1049, against 965 and 927
• Infographic factuality: 0.521, against 0.179 and 0.265, almost three times the Nano Banana 2 score

Google also tested the model with thinking off, and that version still scored ahead of Nano Banana 2 in every category on the list.

What you can make with Nano Banana 2.1

Posters and diagrams with real text. Clear text for posters and intricate diagrams is one of the uses Google names first, and localized text rendering means the same layout can carry a headline in several languages. The jump in infographic design and factuality scores makes explainer graphics and data visuals the obvious first test.

Product scenes from several shots. Multi-product recontextualization brings several product images into one new scene, and product consistency is one of the editing categories where the new model scored higher than Nano Banana 2. Think a skincare set on one styled shelf, or a full outfit arranged in a single frame.

Character series. Multi-character consistency posted the biggest gain in Google’s tests, which makes the model a strong fit for a comic strip, a mascot campaign or a recurring cast of characters across a run of images.

Try Nano Banana 2.1 in AI Playground

Open AI Playground, pick Nano Banana 2.1, and start with a picture you already have. Ask for one change, then another, and see how far a single conversation takes the image.