Nano Banana 2.1 API — How to Call Google's Image Model

Nano Banana 2.1 is Google's next-gen image model on KeyoAPI's OpenAI-compatible /v1/images/generations endpoint — same Bearer key as chat, about $0.0123 per image, no Google Cloud setup.

Listed price: $0.0123 / image · Live catalog: Model Square · Static list: pricing list.

Get API key Open Model Square Compare API prices

Overview

Nano Banana 2.1 is Google's next-generation image model, served on KeyoAPI through the OpenAI-compatible /v1/images/generations endpoint. If you already call chat, Whisper, or OCR with a KeyoAPI key, you can generate images with the same Bearer token — no Google Cloud project, no AI Studio key, and no billing setup on Google's side.

How to call it. Point your HTTP client or OpenAI SDK at https://www.keyoapi.xyz/v1, set model=nano-banana-2.1, and POST a JSON body with a prompt to /v1/images/generations. The prompt field is required; size (for example 1024x1024) is optional. Responses return a generated image URL, so most OpenAI-compatible client code works after changing the base URL and model string.

Pricing. Nano Banana 2.1 is metered per generated image at about $0.0123 — confirm live on /pricing-list or the Model Square page /model/nano-banana-2.1 before you budget. The same endpoint also serves the wider family: nano-banana-2 (~$0.0205/image) when a pipeline was validated on the previous generation, and nano-banana-pro (~$0.0274/image) when the job leans on stronger prompt adherence and text-in-image accuracy.

Worked cost example (indicative). Generating 500 product mockups a month at $0.0123 is roughly $6.15 before retries; the same volume on nano-banana-pro is about $13.70. Batch jobs that tolerate the cheaper model should default to 2.1 and escalate per image when quality gates fail — the same split-lane pattern KeyoAPI recommends for chat models.

Integration steps. (1) Sign up at /sign-up and mint an API key. (2) Point your client at https://www.keyoapi.xyz/v1 with Bearer auth. (3) POST /v1/images/generations with model=nano-banana-2.1 and a prompt. (4) Log each returned image URL next to the Wallet ledger entry so finance sees actuals, not guesses. (5) Try interactively at /model/nano-banana-2.1 — the Model Square try/buy page for this ID.

When to choose which family member. Use 2.1 for volume: mockups, ad variants, storyboards, and batch hero images. Use pro when typography, layout accuracy, or dense prompt following decides success. Keep nano-banana-2 for pipelines already validated on the previous generation — switching mid-flight costs more in QA than it saves per image.

Prompting notes. Generated images track prompt detail: state subject, style, and composition explicitly. For text inside images, prefer pro and spell the text out in quotes. Respect platform usage policies around real people's likenesses and sensitive content.

Related reading on the same site. Background removal for generated art: /model/RMBG-2.0. Open-vocabulary segmentation: /model/sam3. Upscaling before print: /model/Real-ESRGAN. Full model guide index: /models. Docs: /brand/keyo-docs.html.

Quick start

Base URL: https://www.keyoapi.xyz/v1

Endpoint: POST /v1/images/generations

Model: model=nano-banana-2.1 · prompt required · size optional (e.g. 1024x1024)

curl https://www.keyoapi.xyz/v1/images/generations \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"nano-banana-2.1","prompt":"a red paper boat on calm water at sunset"}'

Auth uses the same Bearer API key as chat. Full notes: Keyo docs.

FAQ

How do I call the Nano Banana 2.1 API?

Use your KeyoAPI Bearer key and POST to /v1/images/generations with model=nano-banana-2.1 and a prompt. The endpoint is OpenAI-compatible, so existing images client code works after changing the base URL and model.

How much does Nano Banana 2.1 cost?

About $0.0123 per generated image on KeyoAPI. The same endpoint serves nano-banana-2 (~$0.0205) and nano-banana-pro (~$0.0274). Confirm live rates on /pricing-list.

Is Nano Banana 2.1 better than Nano Banana Pro?

They trade places by job. 2.1 is the cheaper high-volume lane; pro leans on stronger prompt adherence and text-in-image accuracy. Route bulk generation to 2.1 and escalate per image when quality gates fail.

Do I need a Google AI Studio key?

No. KeyoAPI hosts the model; the same key you use for chat and Whisper calls it.

How do I try it before committing volume?

Create an account at /sign-up, mint a key, and test in the Model Square at /model/nano-banana-2.1.

Related models

RMBG-2.0 API SAM 3 API Real-ESRGAN API

Guide for nano-banana-2.1. Catalog: /pricing · Rates: /pricing-list.