Nano Banana 2.1 API launch and pricing

LaoZhang API now serves Google’s Nano Banana 2.1 (gemini-nano-banana-2.1) at $0.045 per call, the same price for 1K, 2K, and 4K. It updates Nano Banana 2 with better text and infographic layouts and costs $0.01 less per call, so existing Nano Banana 2 code switches by changing the model ID.

Published: October 7, 2026
Last verified: October 7, 2026
Status: Active. The model is available in the default token group.

Key facts

Item	Details
Model ID	`gemini-nano-banana-2.1`
Price	$0.045 per call, whatever the output size, thinking level, or number of references
Price source	Public pricing as of October 7, 2026; see the model and pricing catalog
Token	`default` group, billing mode Usage first (按量优先, recommended) or Per-call (按次计费)
Endpoints	Gemini-native `POST /v1beta/models/gemini-nano-banana-2.1:generateContent`, or OpenAI-compatible `POST /v1/chat/completions`
Output sizes	`1K` (default), `2K`, `4K`; `512` isn’t supported
Aspect ratios	14, the same as Nano Banana 2
Reference images	Up to 14: 10 object images and 4 character images
Thinking levels	`minimal`, `medium` (default), `high`
Search grounding	Web search and image search

What changed from Nano Banana 2

	Nano Banana 2.1	Nano Banana 2
Price per call	$0.045	$0.055
Output sizes	1K to 4K	512px to 4K
Default thinking level	`medium`	`minimal`
Ratios, references, grounding	Same	Same

According to Google’s model page, Nano Banana 2.1 improves on Nano Banana 2 in three areas:

More accurate text rendering and infographic layouts
Consistency for up to 4 characters and 10 objects across multi-turn edits
No more tiling artifacts on wide and panoramic ratios such as 1:4, 4:1, 1:8, and 8:1

Who should care

New projects: start with Nano Banana 2.1. Every size from 1K to 4K costs $0.045 per call.
Existing Nano Banana 2 integrations: nothing changes. gemini-3.1-flash-image stays available at the same price. Switch when you want the lower price or better text.
512px output: stay on Nano Banana 2; Nano Banana 2.1 doesn’t accept 512.
1K drafts at the lowest price: Nano Banana 2 Lite, at $0.025 per call, is still the cheapest model in the family.

What to do

Check your token

Use a default-group token with the Usage first (recommended) or Per-call billing mode. Any token that already calls Nano Banana 2 can call Nano Banana 2.1.

Change the model ID

On the Gemini-native route, put gemini-nano-banana-2.1 in the URL. On the OpenAI-compatible route, set model. The request body and response parsing stay the same.

Check two settings

Change any "imageSize": "512" to 1K, or the request returns HTTP 400. Without a thinking level, the model thinks at medium; for faster images, send "thinkingConfig": {"thinkingLevel": "minimal"}.

Request example

Generate a 16:9 image at 2K and save it with jq (version 1.6 or later):

export LAOZHANG_API_KEY="sk-..."  # replace with your LaoZhang API key

curl --fail-with-body --max-time 300 \
  "https://api.laozhang.ai/v1beta/models/gemini-nano-banana-2.1:generateContent" \
  -H "Authorization: Bearer $LAOZHANG_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "contents": [{
      "role": "user",
      "parts": [{"text": "Generate an image of a corner cafe after rain at dusk, warm light on the wet pavement, with a neon sign that reads \"BANANA CAFE\"."}]
    }],
    "generationConfig": {
      "responseModalities": ["IMAGE"],
      "imageConfig": {"aspectRatio": "16:9", "imageSize": "2K"}
    }
  }' \
  -o nb21.json

jq -r '[.candidates[0].content.parts[] | select(.inlineData and (.thought | not))][-1].inlineData.data' nb21.json \
  | base64 --decode > nb21.jpg

Images usually come back as JPEG; read inlineData.mimeType to be sure of the format.
For editing, multiple references, search grounding, and the OpenAI-compatible route, see the Nano Banana 2.1 guide.

FAQ

How much cheaper is Nano Banana 2.1 than Nano Banana 2?

$0.01 per call: Nano Banana 2.1 costs $0.045 and Nano Banana 2 costs $0.055. Both bill per call, so a 1K and a 4K image cost the same. Call logs show the actual charge.

Is Nano Banana 2 being retired?

Not with this launch. gemini-3.1-flash-image stays available at the same price, and any shutdown would get its own announcement.

Does a 4K image also cost $0.045?

Yes. On LaoZhang API, Nano Banana models bill per call only, so neither output size nor thinking level changes the price. Google bills by tokens, about $0.113 for a 4K image, according to its pricing page.

Why does `512` return HTTP 400?

Nano Banana 2.1 accepts only 1K, 2K, and 4K, and requests rejected for invalid parameters aren’t charged. Use Nano Banana 2 for 512px images.

Can I call it through the OpenAI-compatible route?

Yes, set model to gemini-nano-banana-2.1. That route can’t set the ratio or size: it returns a 1K image in a frame the model picks from the prompt. For a fixed ratio or 2K and 4K, use the Gemini-native route.

Nano Banana 2.1 guide: parameters, generation and editing examples, errors, and billing
Nano Banana overview: compare the five models, the two protocols, and billing rules
Choose an image API: compare with GPT Image, Seedream, and FLUX.2
Model and pricing catalog: current prices and token groups
Console pricing: sign in to see the models and prices for your tokens
Google: Gemini Nano Banana 2.1: official capabilities and limits
Google: Gemini API pricing: Google’s token-based prices

​Key facts

​What changed from Nano Banana 2

​Who should care

​What to do

​Request example

​FAQ

​How much cheaper is Nano Banana 2.1 than Nano Banana 2?

​Is Nano Banana 2 being retired?

​Does a 4K image also cost $0.045?

​Why does 512 return HTTP 400?

​Can I call it through the OpenAI-compatible route?

​Sources and related pages