Google Nano Banana 2.1 Image Generator

The upgrade to Nano Banana 2: four people in one shot, every face recognizable, and 8:1 panoramas in one pass with no seams. Every image on this page was made by Nano Banana 2.1

Covers, posters and picture books with cleaner layout and color

In Google's blind tests, 2.1 scores 60 points higher than Nano Banana 2 for text-to-image. All of these were made by 2.1, open one to remix the prompt

Four people in one shot, every face matches

Multi-character consistency is 2.1's biggest gain: 1106 in Google's blind tests, up from 978. Up to 4 people plus 10 objects

  • 4 people, a corgi and 9 objects

    14 reference images become one rooftop picnic. Zoom in and every person and object matches its reference

    Try it with your photos

Extreme aspect ratios, no stitching

Nano Banana 2 left seams in 1:4 to 8:1 images at 2K and 4K. 2.1 fixes that, so light and perspective stay continuous across the frame

An 8:1 banner, or a group shot of four

See how many credits each image costs before you submit

What is Nano Banana 2.1?

Nano Banana 2.1 is Google's image generation and editing model, released on October 6, 2026 as the upgrade to Nano Banana 2. Google says it improves image quality, prompt following, character consistency across multi-turn edits and text rendering, and fixes seam artifacts in 1:4, 4:1, 1:8 and 8:1 images at 2K and 4K. On Pilio you can generate from text or edit with up to 14 reference images, with 1K, 2K or 4K output and no API setup
  • Multi-character consistency: 1106 in Google's blind tests, up from 978
  • Up to 14 reference images: 4 people plus 10 objects
  • 1:8 to 8:1 panoramas with seam artifacts fixed
  • 4.67 credits per 1K image, 10.5 per 4K image
Two reference images, a woman and a corgi, become one autumn street in Kyoto where she walks the corgi, both matching their references

Nano Banana 2.1, Nano Banana Pro or GPT Image 2.5

Pick 2.1 for group shots and extreme aspect ratios; pick GPT Image 2.5 for long text and dense layouts

Nano Banana 2.1

Best for
Group shots, multi-turn edits, banners and long images
Credits per image (1K / 2K / 4K)
4.67 / 7 / 10.5
Aspect ratios
Standard ratios plus 1:4 to 8:1
Reference images
Up to 14 (4 people + 10 objects)
Tiers
Single tier
Output resolution
1K / 2K / 4K

Nano Banana Pro

Best for
Lots of 4K finals
Credits per image (1K / 2K / 4K)
7 / 7 / 9.34
Aspect ratios
Standard ratios, up to 21:9
Reference images
Up to 14
Tiers
Single tier
Output resolution
1K / 2K / 4K

GPT Image 2.5

Best for
Text-heavy posters, UI mockups, refined layouts
Credits per image (1K / 2K / 4K)
2 / 7 / 11
Aspect ratios
Standard ratios
Reference images
Up to 16
Tiers
Flare for speed, Sunburst for detail
Output resolution
1K / 2K / 4K

Not sure? Start with Nano Banana 2.1 and switch to GPT Image 2.5 for long text and dense layouts

A lighthouse, a corgi and a sunset, added in three passes

Across multiple edits, what you already made stays put

  1. Draw the lighthouse

    Name the subject, setting and style, and put any text in quotes

  2. Add a corgi

    Upload a pet photo so it sits in the scene as is, with the lighthouse unchanged

  3. Switch to sunset, then finish

    Change one thing at a time, like "make it sunset", then download at a higher resolution

Nano Banana 2.1 FAQ

What is Nano Banana 2.1?
Nano Banana 2.1 is Google's image generation and editing model, released on October 6, 2026 as the upgrade to Nano Banana 2. Its model ID is gemini-nano-banana-2.1.
What is the difference between Nano Banana 2.1 and Nano Banana 2?
In Google's blind tests, 2.1 scores 1050 for text-to-image, 1048 for infographics, 1026 for editing and 1106 for multi-character consistency, versus 990, 961, 938 and 978 for Nano Banana 2. It also fixes seams in extra-wide and extra-tall images. The trade-off: more thinking by default, slightly slower renders and more credits per image.
Nano Banana 2.1 or Nano Banana Pro?
Pick 2.1 for group shots, banners and long images; it is also cheaper than Pro at 1K and the same price at 2K. For lots of 4K finals, Pro costs fewer credits: 9.34 per 4K image versus 10.5 for 2.1. You can switch between them anytime on Pilio.
How many people can I put in one image?
Up to 4 people and 10 objects, with up to 14 reference images in total. Say which reference belongs to which person and which features must stay the same.
Which resolutions and aspect ratios are supported?
Output is 1K, 2K or 4K. Besides standard ratios, 2.1 supports 1:4, 4:1, 1:8 and 8:1 for banners, website headers and long posters.
Can I use Nano Banana 2.1 for free?
You can try two images. New accounts get 12 credits after first verification, enough for two 1K images on 2.1 (4.67 credits each). After that, buy a credit pack or subscribe. The credits for each image are shown before you submit.
Can Nano Banana 2.1 search before drawing?
Google's 2.1 supports search, but on Pilio search-before-drawing is currently open only for Nano Banana 2, not 2.1.
Do the images have a watermark?
Images generated by Google carry SynthID, an invisible digital watermark embedded in the pixels that does not affect how the image looks.
Do I need an API key or a Google account?
No. Pilio runs Nano Banana 2.1 in your browser. Sign in to Pilio and write a prompt or upload references to generate.