Little illustrations for digital journals
petalumi.

Nano Banana 2: Google's Fast Image Model, in Official Examples

Nano Banana 2 (Gemini 3.1 Flash Image) is Google's fast, high-quality image model. See Google's official examples and prompts, plus specs, pricing and how it compares with Nano Banana Pro and Lite.

3 min read

Nano Banana 2 is Google's name for Gemini 3.1 Flash Image. Google describes it as bringing "the advanced world knowledge, quality and reasoning you love in Nano Banana Pro, at lightning-fast speed." Every image on this page is an official example that Google published with the model, shown with the prompt Google used.

Real-world knowledge, visualized

Nano Banana 2 draws on Gemini's knowledge of the world, so it can explain an idea instead of only illustrating it. Google uses it to turn a topic into an infographic, a diagram or a labeled comparison.

A flat-lay infographic of the water cycle built from paper clouds, cotton balls and hand-drawn arrows
Prompt: “High-quality flat lay photography creating a DIY infographic that simply explains how the water cycle works…” Image: Google
A three-panel comic infographic comparing cumulus, stratus and cirrus clouds
Prompt: “Triptych infographic comparing three types of clouds: Cumulus, Stratus, and Cirrus. Each panel shows the cloud type in a dramatic sky with a bold label. High-contrast comic style. AR: 16:9” Image: Google

Grounded in what is real

The model can search for visual references before it draws. Google asked it to look up a real museum first, then paint it in a new style.

Museum Clos Lucé painted in bright Synthetic Cubism, compared with Nano Banana Pro
Prompt: “Create an image of Museum Clos Lucé. In the style of bright colored Synthetic Cubism. No text. Your plan is to first search for visual references, and generate after. Aspect ratio 16:9” Image: Google

Words you can read, in any language

Google says Nano Banana 2 generates "accurate, legible text for marketing mockups or greeting cards," and can translate and localize the text inside an image.

A small illustrated wildlife sign, shown in English and then localized to an Indian setting with Hindi text
Prompts: “…Delicate script below reads: ‘Native Wildlife: Please Observe from a Distance.’…” then “Take this concept and localize it to an Indian setting, including translation of all the text to Hindi.” Image: Google

For stickers, that covers planner headers, labels, date tabs and short quotes. The sticker platform KLIPY uses the same feature:

Gemini 3.1 Flash Image's Precision Text Rendering allows KLIPY to create accurate copy directly in meme-style assets, stickers, and emojis.

KLIPY, quoted by Google DeepMind

The same characters, every time

According to Google, the model can "maintain character resemblance of up to five characters and the fidelity of up to 14 objects in a single workflow." A mascot can stay recognizable across a whole series.

Fourteen felt-textured farm characters and objects, then all of them together in one farm scene
Prompt: “Create an image of these 14 characters and items having fun at the farm… It is strictly important to keep identity consistent of all the 14 characters and items.” Image: Google
Three fluffy characters building a treehouse across a six-part story
Prompt: “Create a funny 6 part story with these 3 fluffy friends building a tree house… Keep the attire and identity consistent of all 3 characters…” Image: Google
A baby kangaroo in the same hat on a boat, in a café, on a hotel bed and in a bouncy castle
Prompts: “35mm soft blur, vibrant colors, soft light. Sailing at the ocean this baby kangaroo wearing a lilac raincoat…” and three more scenes with the same hat. Image: Google

One visual, every size

Google calls these "production-ready specs": aspect ratios from square to 1:8 strips, and resolutions from 512px to 4K. On Gemini Enterprise Agent Platform, 4K output is still in preview.

The same pop-art portrait composed for many different aspect ratios
Prompt: “Cinematic still, evoking a vibrant, dreamlike quality often found in highly stylized musical dramas…” rendered in several aspect ratios. Image: Google

Photographic quality at Flash speed

Google says Nano Banana 2 "delivers vibrant lighting, richer textures and sharper details, maintaining high-quality aesthetics at the speed expected from Flash."

A moody aerial photograph of a misty green valley with a winding river
Prompt: “This aerial shot captures a dramatic, misty landscape, likely a valley or glen, characterized by rolling, verdant hills and a winding river or loch…” Image: Google

Specs

Nano Banana 2
Model IDgemini-3.1-flash-image
ReleasedFebruary 26, 2026; generally available on Gemini Enterprise Agent Platform since May 28, 2026
InputText, images, video
OutputImages and text
Resolutions512, 1K, 2K, 4K
Aspect ratios15, including 1:8, 8:1, 21:9 and 9:21
Reference images per promptUp to 14
Context window131,072 tokens

Pricing

On the global endpoint of Gemini Enterprise Agent Platform, Google bills image output at $60 per million tokens:

ResolutionOutput tokensApprox. price per image
512747$0.045
1K1,120$0.067
2K1,680$0.101
4K2,520$0.151

Text and image input costs $0.50 per million tokens. These are Google's prices as of October 2, 2026.

Nano Banana 2, Pro or Lite?

If you need…Choose
Fast iteration with high qualityNano Banana 2
Maximum factual accuracy and complex compositionsNano Banana Pro
The lowest cost for 1K draftsNano Banana 2 Lite

Frequently asked questions

Is Nano Banana 2 the same as Gemini 3.1 Flash Image?

Yes. Nano Banana 2 is the product name; gemini-3.1-flash-image is the model ID developers use.

What sizes can it make?

Resolutions of 512, 1K, 2K and 4K, in 15 aspect ratios from 1:1 to 1:8 and 8:1.

Can it put text inside an image?

Yes. Google highlights accurate, legible text and in-image translation as core features.

How can I tell an image was made with AI?

Every image carries Google's invisible SynthID watermark and supports C2PA Content Credentials.

Sources

  1. google-blogNano Banana 2: Combining Pro capabilities with lightning-fast speed — by Google
  2. google-blogBuild with Nano Banana 2, our best image generation and editing model — by Google
  3. google-deepmindGemini 3.1 Flash Image — by Google DeepMind
  4. google-deepmindGemini 3.1 Flash Image model card — by Google DeepMind
  5. google-aiGemini 3.1 Flash Image (Gemini API) — by Google AI for Developers
  6. google-cloudGemini 3.1 Flash Image (Nano Banana 2) model card — by Google Cloud
  7. google-cloudGemini Enterprise Agent Platform pricing — by Google Cloud