Nano Banana API (October 2026): Nano Banana 2.1, 2 Lite and Pro Pricing, API Key and Python Examples

Prices and model details below are from Google's Gemini API documentation on October 17, 2026. We ran the code examples against a test endpoint and recorded the real error messages.

The Nano Banana models

All generated images carry Google's invisible SynthID watermark.

Pricing

Google bills image output by tokens and publishes the equivalent price per image. Standard and Batch prices per image, in USD:

Notes:

• Nano Banana 2.1 is half the price of Nano Banana 2 at 1K and 2K while improving quality, so there's little reason to stay on 2.
• Pro at 2K costs the same as at 1K, so if you use Pro, ask for 2K.
• Text and image inputs are billed too, but they're small: Nano Banana 2.1 charges $1.50 per million input tokens. Its "thinking" is billed as text output at $7.50 per million tokens.
• No free tier: the pricing page marks every Nano Banana model "Not available" on the free tier.

Example: 1,000 product shots at 1K with Nano Banana 2.1 cost about $33.60, or $16.80 through the Batch API. 100 4K hero images with Nano Banana Pro cost about $24.

Compared with other image APIs

Quality differs as much as price, so generate the same prompts on two or three models before committing. See our Grok API page for Grok Imagine details.

Get an API key

1. Sign in to Google AI Studio and create an API key under Dashboard → API Keys. New users get a default project and key automatically.
2. Enable billing. On the Projects page, click Set up billing for your project. New accounts prepay at least $5. When the prepaid balance reaches zero, requests fail with HTTP 402.
3. Save the key as GEMINI_API_KEY. The official SDK reads it automatically.

Since May 28, 2026, keys created in AI Studio are authorization keys bound to a service account, and Google rejects unrestricted standard API keys. If you created an old-style key in the Google Cloud console, add an API restriction or create a new key in AI Studio. Also check Google's available regions list.

Interactions API or generateContent?

Google's image generation guide now uses the Interactions API, which became generally available in June 2026 and is recommended for new projects. The older generateContent API is now labelled legacy but remains fully supported, and two things still need it:

• The Batch API (the 50% discount) isn't available through Interactions yet.
• Most third-party gateways implement generateContent but not yet /v1beta/interactions.

Generate an image with the Interactions API

import base64
from google import genai

client = genai.Client()  # reads GEMINI_API_KEY
interaction = client.interactions.create(
    model="gemini-nano-banana-2.1",
    input="A glossy magazine cover with the words NANO BANANA in a serif font",
)
with open("cover.png", "wb") as f:
    f.write(base64.b64decode(interaction.output_image.data))

Generate an image with generateContent

from google import genai
from google.genai import types

client = genai.Client()  # reads GEMINI_API_KEY
response = client.models.generate_content(
    model="gemini-nano-banana-2.1",
    contents="A glossy magazine cover with the words NANO BANANA in a serif font",
    config=types.GenerateContentConfig(
        response_modalities=["TEXT", "IMAGE"],
        image_config=types.ImageConfig(aspect_ratio="3:4", image_size="2K"),
    ),
)
for part in response.parts:
    if part.inline_data is not None:
        part.as_image().save("cover.png")
    elif part.text:
        print(part.text)

Edit an image

Pass the source image along with the instruction in contents, for example contents=[Image.open("photo.png"), "Replace the background with a beach at sunset"]. Nano Banana 2.1 accepts up to 14 reference images and keeps up to 4 characters and 10 objects consistent across them, which makes it useful for product shots and character series. For several rounds of edits on the same image, use a chat session so the model keeps the previous result.

Options worth knowing

• Aspect ratios: 1:1, 2:3, 3:2, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9 and 21:9, plus panoramic 1:4, 4:1, 1:8 and 8:1 on Nano Banana 2 and 2.1.
• Resolution (image_size): 1K, 2K or 4K. Nano Banana 2 also offers 0.5K; 2 Lite supports 1K only.
• Grounding with Google Search: Nano Banana 2 and 2.1 can look up web and image search results to draw real places, products or animals more accurately.
• Thinking level on 2.1: minimal, medium (default) or high. Higher levels cost more text-output tokens.
• Interleaved output: ask for an illustrated story or a step-by-step guide and you get text and images in one response.

Errors you'll actually see

Using Nano Banana through a gateway

If you already use another model provider's gateway, it may serve Nano Banana as well. api.foo, the service that publishes these docs, serves Gemini image models through the generateContent format: set http_options=types.HttpOptions(base_url="https://api.foo") in genai.Client and keep the rest of the generateContent example above. We confirmed with a test key that requests reach /v1beta/models/gemini-nano-banana-2.1:generateContent. It doesn't support the Interactions API yet, which returns 404. Its exact model names and per-image prices are on the api.foo pricing page. Some Gemini image model names there include a -preview suffix.

Related

• Grok API: pricing, Imagine image and video
• OpenAI API alternatives compared
• Free LLM APIs in 2026
• Install Gemini CLI

Sources

Gemini API documentation, checked October 17, 2026: Nano Banana image generation (Interactions), image generation with generateContent, Interactions API overview, pricing, model pages for Nano Banana 2.1, 2 Lite, 2 and Pro, API keys, billing and rate limits. OpenAI and xAI image prices from their pricing pages on the same date.