Image
Google

Nano Banana

Google's fast, affordable image model for text-to-image and reference-based editing, with flat per-image pricing.

From 20 HGcoins / generation·pay per generation, no subscription
Generated with Nano Banana
Generated with Nano Banana
Examples

Made with Nano Banana

Sample outputs. Open in Studio to generate your own.

What it's for

Where Nano Banana shines

Text to image

Generate a brand-new image from a text prompt alone, with no reference image required.

Photo editing

Attach up to five reference images to switch into edit mode and reshape existing visuals.

Batch variations

Produce up to eight images in one request to compare multiple takes side by side.

Strengths

  • Text-to-image generation that needs no reference image to get started
  • Image editing mode that activates when you add up to five reference images
  • Wide aspect-ratio choice including an Auto option, from 1:1 to 21:9
  • Batch generation of up to eight images per request for fast variation exploration
  • Flat, predictable per-image pricing that scales only with batch count
  • Standard PNG and JPG outputs that drop straight into any workflow

Trade-offs

  • Image-only model with no video, audio, or duration output
  • No resolution selector is exposed, so you cannot choose the output resolution in HexGen
  • Reference images are capped at five per request
Specs

At a glance

Type
Image
Vendor
Google
Modes
Text-to-image, Image editing
Aspect ratios
Auto, 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3, 5:4, 4:5, 21:9
Reference images
Supported (up to 5)
Output
PNG, JPG
Pricing
Flat per image, scaled by batch count

About Nano Banana

Nano Banana is Google's image generation model, built for fast, general-purpose work at an affordable, predictable price. Start from a text prompt alone to create an image from scratch, or attach reference images to switch into editing mode and reshape what you already have.

The model handles a broad set of aspect ratios, from square 1:1 to widescreen 21:9, plus an Auto option that lets the model choose the best fit for your prompt. You can attach up to five reference images to guide an edit, and generate a batch of up to eight images in a single request to explore variations quickly.

Outputs arrive as standard PNG or JPG files, and pricing stays simple: a flat rate per image that scales only with how many images you generate. On HexGen, you write your prompt, pick an aspect ratio, optionally add reference images, set your batch count, and run it.

Prompt ideas

Starting points

Copy, tweak, and run. Good prompts get you most of the way there.

A cozy reading nook by a rain-streaked window, warm lamplight, stacked books and a steaming mug, soft natural color, 4:5 aspect ratio.

Product photo of a matte ceramic coffee cup on a marble counter, clean studio lighting, soft shadows, 1:1 square.

Using the attached reference, change the background to a sunlit garden and keep the subject and pose unchanged.

Pricing
20
HGcoins / generation · ≈ $0.02

Pay only for what you render. 1 USD = 1,000 HGcoins. HGcoins never expire and failed runs refund automatically.

Compare

Nano Banana vs other models

Nano Banana is Google's affordable, general-purpose image tier with simple flat pricing. Here is how it sits next to other image models in the catalog.

Nano Banana vs other models
ModelQualitySpeedCostChoose it when
Nano Banana
This
Google
Great
Fastest
Lower cost
Pick Nano Banana when you want fast text-to-image and reference editing at the lowest, most predictable per-image cost.
Best
Fast
Higher cost
Step up to the Pro tier when you need a higher-end Google image variant and can spend more per image.
ByteDance
Best
Fast
Mid cost
Choose Seedream when you want an alternative image model from ByteDance for your generations.
Bottom line: pick Nano Banana when pick nano banana when you want fast text-to-image and reference editing at the lowest, most predictable per-image cost.. Otherwise one of the models above will fit better. Tap a row to compare.

Frequently asked questions

Nano Banana is Google's image generation model. It supports text-to-image generation and image editing with reference images, and outputs standard PNG and JPG files.