GOOGLE IMAGE MODEL

Nano Banana 2.1

Characters, settings and illustration: explore our test images.

Choose 1K, 2K or 4K output. Supported reasoning and image references do not add a customer surcharge.

Nano Banana 2.1 image workflow
1K / 2K / 4K
MINIMAL / MEDIUM / HIGH

Nano Banana 2.1 image workflow

Text-to-image and image editing through Google Vertex AI.

View pricing

1K, 2K and 4K

Choose the output size and one of fourteen documented aspect ratios. No seed, sampling or native mask control is offered.

Owned image references

Attach one to fourteen owned images in the intended order and explain what each should preserve or change.

Reasoning

Choose minimal, medium or high. Reasoning usage varies and is separate from image output tokens.

Exact quote before generation

Available in MaxVideoAI through Google Vertex AI.

Exact quote before generation

Choose 1K, 2K or 4K output. Supported reasoning and image references do not add a customer surcharge.

View pricing

Standard preview

$0.07

1k

2K still

$0.11

2k

4K still

$0.23

4k

Generation example: 1 image · 1k · Text to image — $0.07

Your current quote is confirmed before generation. Supplier text and reasoning costs do not change the agreed image price.

14 images generated with Nano Banana 2.1

Nano Banana 2.1 · Text to image · 1K · 16:9

These test images were generated from text with Nano Banana 2.1 via Google Vertex AI. Nano Banana 2.1 is a separate model from Nano Banana 2.

Open image workspace

Explore Kling 4.0 videos

Nano Banana 2 remains available

Nano Banana 2 retains its own model identity, settings and history.

View Nano Banana 2

Describe the image

Text-to-image

Describe the subject, action, setting, framing and lighting. Select the aspect ratio, size and reasoning level.

Owned image references

Attach one to fourteen owned images in the intended order and explain what each should preserve or change.

The blue dress

The original prompt used for the Nano Banana 2.1 character still.

Open image workspace

The blue dress

1K · 16:9 · MEDIUM

Original full-body cinematic fashion portrait, horizontal 16:9. Iris, a tall woman aged58, sharply cut silver bob, strong nose, fine eye wrinkles, composed professional expression. A sculptural midnight-blue heavy silk gown with a high asymmetric collar, one plain silver cuff, exactly two simple black heels worn on her feet. Narrow runway end under hard white light; charcoal curtain behind. Photographers stay outside frame. Rich real silk texture, natural mature skin, restrained contrast, poised weight on one leg, no exaggerated beauty retouch. Both shoes and complete dress hem visible. No celebrity likeness, brands, captions, collage or text. This is an identity and wardrobe reference before she leaves the runway. Production language: English for any permitted diegetic writing or visible text. This does not override the request for no typography or text; do not add captions, titles or labels.

Tips and boundaries

Best practices, common fixes, and important limitations to help you get the strongest results with Nano Banana 2.1.

What works best

  • Choose 1K, 2K or 4K output. Supported reasoning and image references do not add a customer surcharge.
  • Choose MINIMAL, MEDIUM or HIGH reasoning, fourteen aspect ratios, and up to fourteen image references for edits. Supported options do not add a customer surcharge.

Specifications

Formats, settings and limits to check before creating.

Price / image

4k $0.23/image1k $0.07/image2k $0.11/image

Text-to-Image

Supported

Image-to-Image

Supported

Resolution options

1K / 2K / 4K

Aspect ratios

1:1 / 3:2 / 2:3 / 3:4 / 1:4 / 4:1 / 4:3 / 4:5 / 5:4 / 1:8 / 8:1 / 9:16 / 16:9 / 21:9

Output format

Provider-selected image format

Release date

Oct 2026

Safety & people / likeness

Built-in safeguards and best practices for responsible creation with Nano Banana 2.1.

  • Don’t generate real people or public figures (celebrities, politicians, etc.).
  • No minors, sexual content, hateful content, or graphic violence.
  • Don’t use someone’s likeness without consent.
  • Some prompts and reference images may be blocked — generic characters and scenes are fine.