Nano Banana 2 Lite — Fast 1K image drafts to test your ideas.
Nano Banana 2 Lite is Google’s name for Gemini 3.1 Flash-Lite Image, announced in June 2026 as the fastest model in the Nano Banana family. On Deepnia it is the default image model: it creates or edits 1K images in 14 aspect ratios with up to 10 reference images, which suits testing an idea, comparing drafts or preparing a social post.
The image tool opens with Nano Banana 2 Lite already selected. The exact cost shows before every creation; you sign up when you generate.
9:16
1:1
4:5
16:9
4 of the 14 aspect ratios on Deepnia, all in 1K.
At a glance
Maker
Google
Google name
Gemini 3.1 Flash-Lite Image
Resolution
1K, in every aspect ratio
Aspect ratios
14 + Auto (9:16 by default)
References
Up to 10 images
Images per generation
1
Prompt
Up to 20,000 characters
What Nano Banana 2 Lite does best
Built for speed
Google presents Nano Banana 2 Lite as the fastest and most efficient model in the Nano Banana family, with text-to-image results in about 4 seconds. On Deepnia the wait also depends on the queue, but it is the natural place to start testing an idea.
Variations in one click
Each generation gives 1 image: click “Regenerate” to run the same request again and compare versions, then refine the one you like. Handy when choosing a social visual, a thumbnail or a logo idea.
14 aspect ratios, from square to banner
Square (1:1) for feeds, 9:16 for stories, 16:9 for thumbnails, 21:9 for banners: 14 fixed aspect ratios plus Auto. Every image comes out in 1K.
Simple edits from your photos
Add up to 10 reference images and the tool switches to edit mode: change a color, a background or a detail. For a composition that combines many references, switch to Nano Banana 2.
Readable text and consistent characters
Google says Lite keeps reliable prompt adherence, strong character consistency and legible in-image text despite its speed. Always proofread the spelling before you publish.
How to use Nano Banana 2 Lite on Deepnia
Open the image tool
The “Try Nano Banana 2 Lite” button opens Deepnia’s image tool with the model already selected. It is also the tool’s default model.
Add a photo if you need one
To edit an image, click “Add an image” under “References” (up to 10, JPG, PNG or WebP up to 10 MB each): the tool switches to “Edit mode”.
Describe the image
Write the prompt: subject, setting, light, style and, if any, the exact text to show in quotation marks.
Pick the aspect ratio and generate
Set the “Aspect ratio”, check the cost shown, then click “Generate”, or “Edit” if you added a reference. “Regenerate” runs the same request again for another version.
Nano Banana 2 Lite on Deepnia: what you can set
Inputs
Text only, or text plus reference images
Modes
Generation without references; editing as soon as you add one
Up to 10 images; JPG, PNG or WebP, up to 10 MB each
Images per generation
1
Prompt
Required, up to 20,000 characters
Prompt ideas to try
Copy a prompt, open the tool and adapt it to your project.
Logo for a café
1:1
Text only
Minimalist logo for a neighborhood café called "Grain & Co": a stylized cup made of two coffee beans, bold lines, two colors (warm brown and cream), plain white background. The name "Grain & Co" in rounded, clearly readable letters under the symbol.
Bright product photo of a sage-green stainless steel water bottle on a light wooden counter, blurred monstera leaves in the background, soft morning light. Leave a large empty area at the top of the image for a headline. Catalog photo style, no text in the image.
Vertical story visual: a smiling fitness coach in black sportswear doing a forward lunge in a park at sunrise, golden light, softly blurred background. At the top, the text "Saturday session, 8 a.m." in large, clearly readable white letters.
Video thumbnail for a podcast about entrepreneurship: two people talking face to face behind microphones, a studio with purple and orange lights, warm atmosphere. On the right, the title "Launch your brand" in large, bold yellow letters that are perfectly readable.
Elegant wedding invitation card: a watercolor wreath of eucalyptus and white roses on an ivory background. In the center, the names "Sara & Karim" in gold calligraphy, and below, "June 14" in small capitals. Refined stationery style with plenty of space around the text.
Describe the scene in full sentences rather than a list of keywords: Google recommends this approach for Nano Banana models.
To compare several directions, click “Regenerate”: the same request runs again and gives a new version. Then adjust the prompt of the one you like.
Put the text to display in quotation marks, describe the lettering (bold, handwritten, gold…) and keep it short.
For an edit, say exactly what changes, then add “keep everything else exactly the same”: Google suggests this wording to preserve the style, lighting and composition.
Describe what you want rather than what you don’t: Google suggests “an empty, deserted street” rather than “no cars”.
Once the idea works in 1K, run the same prompt with Nano Banana 2 for a 2K or 4K version.
Good to know
Pick 9:16 for stories, Reels and TikTok, 4:5 or 1:1 for feed posts, 16:9 for YouTube thumbnails and 21:9 for a website banner.
To keep a character or product identical from one image to the next, add the same photo as a reference in every creation and describe it with the same words.
For a clean edit, change one element at a time: a color, a background or an object.
Before you publish, proofread the spelling, translations and infographic figures, and check small faces and fine details.
Nano Banana 2 Lite or another model?
Deepnia’s other image models, to pick the right one for your project.
Exact pixel sizes set for each aspect ratio, with up to 10 references.
Frequently asked questions
What is Nano Banana 2 Lite?
It is Google’s name for Gemini 3.1 Flash-Lite Image, announced on June 30, 2026. Google presents it as the fastest and most efficient model in the Nano Banana family, built for high-volume work.
Why is it selected by default on Deepnia?
It is the image model Deepnia preselects in the tool; you can switch to another one at any time under “Model”. Google designed it for fast, high-volume creation, which makes it a good starting point.
What resolution do I get?
Every image comes out in 1K, in every aspect ratio. For a final version in 2K or 4K, run the same prompt with Nano Banana 2 or Nano Banana Pro.
How many reference images can I add?
Up to 10, JPG, PNG or WebP up to 10 MB each. One photo is enough for a simple edit; for a composition that combines many images, Nano Banana 2 accepts up to 14.
Is it really fast?
Google cites about 4 seconds for a text-to-image result. On Deepnia, the actual wait also depends on the queue.
Which aspect ratios are available?
Auto and 14 fixed aspect ratios: 1:1, 16:9, 9:16, 2:3, 3:2, 4:3, 3:4, 4:5, 5:4, 21:9, 1:4, 4:1, 1:8, 8:1. 9:16 is selected by default.
Can it write text in languages other than English?
Google announces legible in-image text and lists French, Spanish, Arabic, Hindi and Japanese, among others, as languages where its image models perform best. Always proofread the result, as mistakes remain possible.
Are the images marked as AI-generated?
Yes: Google embeds an invisible SynthID watermark in every image this model creates.
Ready to try Nano Banana 2 Lite?
The image tool opens with Nano Banana 2 Lite already selected. The exact cost shows before every creation; you sign up when you generate.