Nano Banana 2.1, a model for AI image generation and editing, is now available in AI OGO: 4K output, up to 10 source photos, 14 aspect ratios and prompts up to 20,000 characters for product photos, ads and infographics.
Nano Banana 2.1, an updated model for AI image generation and editing, is now available in AI OGO. You can write a text prompt and get a finished image from scratch, or upload your own photos and ask the model to change the background, clothing, objects, style or composition. Characters stay recognizable and important details are preserved, so the result still resembles the original instead of turning into a random jumble of pixels.
What's new in Nano Banana 2.1?
This is an update to Nano Banana 2, and public sources point to several notable improvements: better visual quality, more accurate handling of complex instructions, stronger text rendering, more consistent characters and better wide-format images. The developers also tackled a long-standing pain point for image generators: text in panoramic formats often broke apart into a mosaic. The model outputs images at 1K, 2K and 4K and eliminates segmentation artifacts in panoramic images with extreme aspect ratios, up to 1:8 and 8:1. For ad banners and wide covers, that makes a real difference.
The second set of improvements concerns text inside images and infographics. Nano Banana 2.1 produces higher-quality images, follows instructions more closely and handles text and infographics better, rendering lettering and layouts more precisely. If you need a poster with a headline, a diagram with labels or a product card with a price tag right on the image, your chances of getting legible, neatly aligned text on the first try are noticeably higher than before. The third area is working with multiple sources at once. Multi-image composition with support for up to 14 references lets you bring different characters, objects and visual styles from separate photos together in a single scene.
What settings are available in AI OGO?
Resolution and source photos
Here's exactly what the AI OGO generation settings offer. Resolution options are 1K, 2K and 4K; choose based on whether the image is for social media or large-format print. For editing and visual references, you can upload up to 10 source images in JPEG, PNG or WebP, each up to 30 MB. The model returns results in JPG or PNG.
Aspect ratios and prompt length
There are fourteen aspect ratios: 1:1, 2:3, 3:2, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9, 21:9, 1:4, 4:1, 1:8 and 8:1, plus an automatic option if you'd rather not pick one manually. Prompts can be long and detailed, up to 20,000 characters. That's useful for complex scenes with many details you want to spell out explicitly rather than leave the model to guess.
What can you use the model for?
In practice, Nano Banana 2.1 works well for product photos in marketplace listings, ad layouts, article illustrations, event posters and educational infographics with labels. Multi-reference mode is especially handy when you need to combine a character from one photo, an object from another and a style from a third in a single composition while keeping each element recognizable.
How to try Nano Banana 2.1
You can try the model today in the AI OGO image generation section. If you regularly work with the same characters, save their references in advance under saved characters so you don't have to upload photos again for every new image. For the full list of available models and current generation prices, see the model and pricing catalog.
