Nano Banana 2.1
Nano Banana 2.1 is Google DeepMind's update to Nano Banana 2 (Gemini 3.1 Flash Image), released on 6 October 2026 as gemini-nano-banana-2.1. Google's model card says it is built on Gemini 3.6 Flash. It keeps the Flash tier's speed and 1K, 2K and 4K sizes, drops the 512px size, adds a medium thinking level as the default, and on the Gemini API costs less per image than the model it replaces, which Google deprecated the same day. In the Gemini app it is the image model for Google AI Plus, Pro and Ultra subscribers; free accounts stay on Nano Banana 2.
Anvisha Pai, Co-founder & CEO, Voyager
Google DeepMind · released
›What we asked for
A self-portrait of Nano Banana 2.1, Google's updated everyday AI image model, as a bold flat vector illustration: a cheerful banana character mid-stride with speed lines, wearing a small badge reading '2.1', limited palette of yellow, teal and off-white, thick clean shapes, no outlines, poster composition.
google/nano-banana-2.1
aspect_ratio: 4:3, resolution: 1K, thinking_level: medium, output_format: png, num_images: 1
returned 1200 by 896 pixels
At a glance
- Sizes
- 1K, 2K and 4K in fourteen fixed shapes, including 8:1 and 1:8 strips. The 512px size from Nano Banana 2 is gone.
- Thinking
- On by default. Three levels, minimal, medium and high, with medium as the default; Nano Banana 2 had minimal and high and defaulted to minimal.
- Reference photos
- Up to 14 images per request on Google's side: high fidelity for up to 10 objects and resemblance for up to 4 characters.
- Video and PDF context
- The edit endpoint can take a video or a PDF as context, for thumbnails or summary graphics.
- No seed
- Vertex AI rejects seed, temperature, topP and topK for this model, so a result cannot be pinned and replayed.
Other Nano Banana models: Nano Banana 2, Nano Banana 2 Lite, Nano Banana Pro. Compare all 4.
Anvisha's take
On our twelve prompts 2.1 mostly fixed the habits I warned about on the Nano Banana 2 page. The banner came back with its copy space clean instead of a painted 'Learn More' button, the product shot and its edit carried no embossed brand name, and the poster kept to the three lines asked for. It is not a different model in kind: set text was already right on 2, the woodblock print still arrives signed, and the chart still gains an axis unit nobody gave. It ran at the default medium thinking and took 9 to 20 seconds a picture, a little slower than 2 on its worst prompt. At about half the API price per 1K image, I would move to it.
Nano Banana 2.1 real output: the same 12 prompts every model gets
Every image model on this site gets the same 12 prompts, so you can compare like with like. These are the pictures Nano Banana 2.1 returned, one try each, nothing re-rolled or retouched. Generated on . The 12 pictures took 9 to 20 seconds each and cost $0.484 in total.
Spelling, kerning and hierarchy of three lines of set text.
›What we asked for
A concert poster that reads 'MODA SUMMER SESSIONS' in large bold type, with the subtitle 'Friday 12 June, 8pm, Roundhouse' and a smaller line 'Tickets at the door'. Risograph style, two colours, tangerine and navy, grainy ink texture.
google/nano-banana-2.1
aspect_ratio: 3:4, resolution: 1K, thinking_level: medium, output_format: png, num_images: 1
returned 896 by 1200 pixels
Catalogue lighting, a clean seamless background and a believable contact shadow.
›What we asked for
Studio product photograph of a matte black ceramic pour-over coffee dripper on a seamless white background, soft top light, subtle contact shadow, centred, catalogue style.
google/nano-banana-2.1
aspect_ratio: 1:1, resolution: 1K, thinking_level: medium, output_format: png, num_images: 1
returned 1024 by 1024 pixels
A natural face, correct finger count and hands holding an object.
›What we asked for
Close-up portrait of a woman in her 40s with short grey hair, laughing, holding a red ceramic mug with both hands, natural window light, 85mm lens, shallow depth of field.
google/nano-banana-2.1
aspect_ratio: 3:4, resolution: 1K, thinking_level: medium, output_format: png, num_images: 1
returned 896 by 1200 pixels
A flat single-colour mark with a legible wordmark and no gradients or photo texture.
›What we asked for
A flat vector logo mark for a bicycle repair shop called 'Spoke & Chain': a single-colour icon combining a chain link and a wheel spoke, with the wordmark below in a geometric sans-serif, black on white.
google/nano-banana-2.1
aspect_ratio: 1:1, resolution: 1K, thinking_level: medium, output_format: png, num_images: 1
returned 1024 by 1024 pixels
Many distinct objects and people in one frame without merging or duplicating them.
›What we asked for
A crowded Saturday farmers market seen from slightly above: at least twelve stalls with striped awnings, shoppers carrying tote bags, a dog on a leash, crates of oranges, a musician with a guitar, bunting overhead, late morning light.
google/nano-banana-2.1
aspect_ratio: 4:3, resolution: 1K, thinking_level: medium, output_format: png, num_images: 1
returned 1200 by 896 pixels
A named historical style applied faithfully rather than a generic painterly look.
›What we asked for
A lighthouse on a rocky cliff during a storm, in the style of a Japanese ukiyo-e woodblock print, with visible wood grain texture, flat colour areas and a bokashi gradient sky.
google/nano-banana-2.1
aspect_ratio: 3:4, resolution: 1K, thinking_level: medium, output_format: png, num_images: 1
returned 896 by 1200 pixels
Six labelled bars at the stated heights, in order, with readable axes.
›What we asked for
A clean bar chart titled 'Monthly active users' showing six months, January through June, with the values 120, 150, 180, 170, 220 and 260, labelled axes, white background, one accent colour, no 3D effects.
google/nano-banana-2.1
aspect_ratio: 4:3, resolution: 1K, thinking_level: medium, output_format: png, num_images: 1
returned 1200 by 896 pixels
A 21:9 composition that leaves the left third clear for headline text.
›What we asked for
Website hero banner: a wide panorama of a sunlit modern co-working space with plants, the left third of the frame kept as clear wall space for text, warm neutral palette.
google/nano-banana-2.1
aspect_ratio: 21:9, resolution: 1K, thinking_level: medium, output_format: png, num_images: 1
returned 1584 by 672 pixels
Four numbered steps in the given order with their exact wording, plus an icon each.
›What we asked for
An infographic card titled 'How to brew pour-over' with four numbered steps in this order: 1. Rinse the filter, 2. Bloom for 30 seconds, 3. Pour in slow circles, 4. Wait 3 minutes. A simple icon beside each step, muted palette, clean sans-serif type.
google/nano-banana-2.1
aspect_ratio: 3:4, resolution: 1K, thinking_level: medium, output_format: png, num_images: 1
returned 896 by 1200 pixels
Appetising, menu-grade food with correct textures and no melted anatomy.
›What we asked for
Overhead photograph of a margherita pizza on a wooden board, fresh basil leaves, one slice pulled out with the cheese stretching, restaurant menu photography, natural light.
google/nano-banana-2.1
aspect_ratio: 1:1, resolution: 1K, thinking_level: medium, output_format: png, num_images: 1
returned 1024 by 1024 pixels
Exact counts and left-right placement of five named objects.
›What we asked for
A glass of water on a wooden table with exactly three ice cubes in it, a yellow pencil resting across the rim of the glass, a blue book to the left of the glass and a red apple to the right, photographed from table height.
google/nano-banana-2.1
aspect_ratio: 4:3, resolution: 1K, thinking_level: medium, output_format: png, num_images: 1
returned 1200 by 896 pixels
Change one object's colour and add a part while leaving background, light and framing untouched.
›What we asked for
Change the coffee dripper to a deep forest green and add a small brass handle on its right side. Keep the background, lighting and composition exactly as they are. (applied to the product on white result above)
google/nano-banana-2.1/edit
aspect_ratio: 1:1, resolution: 1K, thinking_level: medium, output_format: png, num_images: 1
returned 1024 by 1024 pixels
How the cost under each picture was worked out: fal's representative $0.040 for a 1K text-to-image request at medium thinking (fal.ai/models/google/nano-banana-2.1, read 2026-10-08); fal bills this endpoint by tokens and does not return the charge.
Compare Nano Banana 2.1
Nano Banana 2.1 pricing
| Where | Unit | Price | Source |
|---|---|---|---|
| Gemini API | 1K image $30 per million image output tokens; a 1K image is 1,120 tokens. Input is $1.50 and text or thinking output $7.50 per million tokens on top, so the real cost per request is a little higher. | $0.0336 | Gemini API pricing, read |
| Gemini API | 2K image | $0.0504 | Gemini API pricing, read |
| Gemini API | 4K image The pricing page counts a 4K image as 3,780 tokens. Google's image generation guide lists 2,520 tokens at 4K for this model, which would be about $0.076; we show the pricing page's figure. | $0.113 | Gemini API pricing, read |
| Gemini API, Batch | 1K image Batch halves every rate: $0.0252 at 2K and $0.0567 at 4K. | $0.0168 | Gemini API pricing, read |
| fal | 1K image Billed by tokens: $35.294 per million image output tokens, $8.823 text and thinking output, $1.764 input. fal's own figure for a short prompt at medium thinking; $0.059 at 2K, $0.134 at 4K, and about $0.004 more for an edit with two 1K inputs. The basis for our sample run. | $0.04 | fal model page, read |
| fal | 4K image | $0.134 | fal model page, read |
Is Nano Banana 2.1 free?
Not free. In the Gemini app Nano Banana 2.1 needs a Google AI plan; free accounts generate with Nano Banana 2. The Gemini API lists no free tier for it.
- •Gemini app, free: image generation runs on Nano Banana 2, not 2.1.
- •Gemini app, Google AI Plus, Pro and Ultra: Nano Banana 2.1, higher generation limits, and the option to redo an image with Nano Banana Pro.
- •Gemini API: the free tier is listed as not available for gemini-nano-banana-2.1.
- •Google Search grounding: 5,000 free search requests a month shared across Gemini 3.x models, then $14 per 1,000.
Where to use Nano Banana 2.1
- Gemini app · The image model on Google AI Plus, Pro and Ultra plans.
- Google AI Studio and Gemini API · Model id gemini-nano-banana-2.1. Pay per token, Batch at half price.
- Gemini Enterprise Agent Platform (Vertex AI) · Same model id, generally available in the global region, pay as you go or provisioned throughput.
- fal · google/nano-banana-2.1 and google/nano-banana-2.1/edit, billed by tokens.
- Voyager · Runs inside Voyager, the agent for creative work. Private preview by waitlist.
Voyager
Run Nano Banana 2.1 inside Voyager
Voyager is an agent for creative work that runs every model in this reference, Nano Banana 2.1 included. Brief it, and it picks the model, makes the thing and hands back something you can edit. Create an account to get started.
What we noticed running Nano Banana 2.1
- •It left the copy space alone. The left third of the 21:9 banner is plain wall, with no button or text painted in, where Nano Banana 2 added a green 'Learn More' button to the same prompt. See the wide aspect ratio picture.
- •No brand leak this time. Neither the black dripper nor its green edit carries the embossed 'V60' that Nano Banana 2 put on both. The edit replaced the ceramic handle with a brass one rather than adding brass beside it, a fair reading of 'add a small brass handle'. See the edit an existing image picture.
- •All three lines correct, in the case they were written, and nothing added: no invented address line. The illustration underneath is an abstract crowd with music notes rather than a venue. See the text in an image picture.
- •Six bars at the stated values, in order, with the title in the prompt's own sentence case. It still invents an axis label, '(Millions)', a unit the prompt never gave. See the chart picture.
- •Steps in order with their exact wording, now numbered and laid out as a two-by-two grid. Each step still carries a line of instructions nobody wrote. See the infographic with steps picture.
- •Style prompts still come signed. The woodblock print has a title cartouche, two red seals and painted signatures. Crop or repaint them before commercial use. See the specific art style picture.
- •The mark combines the two things asked for, a chain link set inside a spoked wheel, and the wordmark is spelled right. It added a 'BICYCLE REPAIR' tagline that was not requested. See the logo mark picture.
- •Book left, apple right, pencil across the rim, and the glass appears to hold the three ice cubes asked for, where Nano Banana 2 put in four. One picture is not proof that counting is fixed. See the hard composition picture.
- •Every picture came back in 9 to 20 seconds at 1K with thinking at the default medium, against 9 to 14 for Nano Banana 2 with thinking off. The slowest was the woodblock print.
- •1K sizes are the same as Nano Banana 2's: 1024 by 1024 at square, 896 by 1200 at 3:4, 1584 by 672 at 21:9.
Nano Banana 2.1 limits and API parameters
| Limit | Value |
|---|---|
| Output size | 1K, 2K or 4K in fourteen fixed shapes. At 1K: 1024 by 1024 square, 896 by 1200 at 3:4, 1584 by 672 at 21:9. No 512px size. |
| Reference images | Up to 14 images per prompt, 7 MB each inline or 30 MB from Cloud Storage. Google rates fidelity for up to 10 objects and resemblance for up to 4 characters. |
| Context | 131,072 input tokens and 32,768 output tokens on Vertex AI; video input up to 10 files or one YouTube URL per prompt. |
| Seed and sampling | seed, topK, topP, temperature and logprobs are not supported; setting any of them returns an error. |
| Known weak spots | Google lists blurry small text at 1K, imperfect character consistency, ink persisting after mask edits, occasional left and right confusion, and poses carried over from an input image. |
| Rate limits | Google meters images per minute per project tier and shows the live numbers in AI Studio rather than publishing a table. Gemini API rate limits, read |
›Every setting, for developers
fal (google/nano-banana-2.1)
fal model page and API reference, read
| Parameter | Type | Values | Default |
|---|---|---|---|
| prompt | string | 3 to 50,000 characters Required. | |
| aspect_ratio | enum | auto, 21:9, 16:9, 3:2, 4:3, 5:4, 1:1, 4:5, 3:4, 2:3, 9:16, 4:1, 1:4, 8:1, 1:8 | auto |
| resolution | enum | 1K, 2K, 4K | 1K |
| num_images | integer | 1 to 4 | 1 |
| output_format | enum | jpeg, png, webp | png |
| thinking_level | enum | minimal, medium, high Thinking is billed as output text tokens. | medium |
| enable_web_search | boolean | true, false No separate fee on fal; search results are billed as input tokens. | false |
| system_prompt | string | Any text | empty |
| safety_tolerance | enum | 1 to 6, 1 is strictest | 4 |
| limit_generations | boolean | true, false Experimental. False lets the model return extra images and text. | true |
| image_urls | list of URLs | Up to 14 images Edit endpoint only, where it is required unless video_url or pdf_url is set. | |
| video_url, pdf_url | URL | A video up to 15 MB or a YouTube URL; a PDF up to 15 MB Edit endpoint only. |
Gemini API (gemini-nano-banana-2.1)
Gemini API image generation guide, read
| Parameter | Type | Values | Default |
|---|---|---|---|
| aspect_ratio | enum | 1:1, 1:4, 1:8, 2:3, 3:2, 3:4, 4:1, 4:3, 4:5, 5:4, 8:1, 9:16, 16:9, 21:9 | |
| image_size | enum | 1K, 2K, 4K Uppercase K required. 512px is not supported on this model. | |
| reference images | inline images | Up to 14 in total: up to 10 objects at high fidelity, up to 4 characters | |
| thinking level | enum | minimal, medium, high | medium |
| grounding | tool | Google Search and Google Image Search |
What Nano Banana 2.1 will not draw
Google's Generative AI Prohibited Use Policy applies on every surface. It bars child sexual abuse material, non-consensual intimate imagery, content that facilitates violent extremism, self-harm or illegal activity, hate speech and harassment, and impersonating a real person without disclosure in order to deceive. There is no adult content mode. fal exposes a safety_tolerance dial from 1 to 6 but it operates inside the same policy. Every image carries an invisible SynthID watermark, and search grounding will not use real photos of people found on the web.
Alternatives to Nano Banana 2.1
Frequently asked questions
What is Nano Banana 2.1?
Nano Banana 2.1 is Google's update to Nano Banana 2, released on 6 October 2026 with the API id gemini-nano-banana-2.1. Google's model card says it is based on Gemini 3.6 Flash. It replaces Nano Banana 2 on the Gemini API, which Google deprecated the same day.
What changed from Nano Banana 2?
Google cites better visual quality, prompt adherence, character consistency over several turns, text rendering and wide panoramic shapes. On the spec sheet: a medium thinking level that is now the default, no 512px size, video and PDF inputs, and a lower API price per image, $0.0336 at 1K against $0.067, read on 8 October 2026.
Is Nano Banana 2.1 free?
No. In the Gemini app it needs a Google AI Plus, Pro or Ultra plan; free accounts generate with Nano Banana 2. The Gemini API has no free tier for it.
Does Nano Banana 2.1 have an API?
Yes. The model id is gemini-nano-banana-2.1 on the Gemini API and on Google's Agent Platform, formerly Vertex AI. fal hosts it as google/nano-banana-2.1, with an edit endpoint at google/nano-banana-2.1/edit.
Can I set a seed?
Not on Google's own platform: Vertex AI returns an error if seed, temperature, topP or topK is set for this model. The same prompt will give a different picture each time.
How this page is made
Prices, limits and settings are read from the linked vendor and host pages on the dates shown. Every picture was generated by us through fal with the request shown under it, and the original files are kept unedited. A generation that failed is shown as a failure. Model and vendor names are used to identify the products; no vendor imagery appears on this page.
