Nano Banana 2.1 or Nano Banana 2? Per-image pricing and differences

Short answer: same endpoint, different price structure

LLMTR (llmtr.com) is a gateway that gives access to language models hosted in Turkey and at global providers through a single OpenAI-compatible API. Nano Banana 2.1 (google/gemini-nano-banana-2.1) and Nano Banana 2 (google/gemini-3.1-flash-image) are both called there with the same API key on the /v1/images/generations endpoint under https://llmtr.com/v1; switching between them means changing only the model value.

The difference is in how they are priced. Using Google's per-image equivalents, a 1K image costs $0.0336 on Nano Banana 2.1 and $0.067 on Nano Banana 2. On the other hand, input tokens ($1.50 vs $0.50) and thinking tokens ($7.50 vs $3) cost more on Nano Banana 2.1. The 512 size is published only for Nano Banana 2. Google describes Nano Banana 2.1 as an update to Nano Banana 2; both models are available on LLMTR.

Nano Banana 2.1 or Nano Banana 2: which one should you choose?

The choice depends on where your workload spends its tokens. If you generate many images from short prompts, the total drops on Nano Banana 2.1, whose per-image equivalents are lower: $0.0336 vs $0.067 at 1K, $0.0504 vs $0.101 at 2K and $0.113 vs $0.151 at 4K. If you send long prompts and many reference images, the input price gap ($1.50 vs $0.50 per 1M tokens) takes a larger share of the total.

If you need the 512 size for small previews, Nano Banana 2 publishes it at $0.045 per image; the sizes published for Nano Banana 2.1 are 1K, 2K and 4K. Quality and style vary from prompt to prompt, so running the same prompts on both models and comparing the results on your own examples gives a more reliable decision than the numbers in the table. An existing integration keeps working unchanged on the google/gemini-3.1-flash-image id.

Price and feature table

Both models are billed by tokens. The per-image equivalents are Google's published conversions; the bill adds prompt, reference and thinking tokens as separate items. The values below are the rows in the LLMTR catalog and match the provider's prices.

• Google, Gemini API pricing · 2026-10-10

Worked example: 1,000 images at 2K

Take a job that generates 1,000 images at 2K, each from a 300-token prompt. On Nano Banana 2.1 the images cost 1,000 x $0.0504 = $50.40 and the prompts 300,000 tokens x $1.50 / 1,000,000 = $0.45, $50.85 in total. On Nano Banana 2 the images cost 1,000 x $0.101 = $101.00 and the prompts 300,000 tokens x $0.50 / 1,000,000 = $0.15, $101.15 in total.

Thinking tokens are not in this calculation and are billed at the text output price on both models. Assuming 1,000 thinking tokens per image, 1,000 images add $7.50 on Nano Banana 2.1 and $3.00 on Nano Banana 2. Edits with reference images raise the input share; compare the real amount of your own requests on the usage page in the Dashboard.

Running the same prompt on both models

For a fair comparison, send the same prompt at the same size to both models. The Python example below asks both models for the same poster as a 1K image at 3:4 and saves each result under the model name.

n is not applied on Nano Banana 2.1; repeat the request for more than one sample. Use 1K, 2K or 4K as the size on 2.1. For a comparison with references, send the same file_ids list to both models; uploading the file once is enough.

Python example that sends the same prompt as a 1K image at 3:4 to both models and saves each image under the model name.

import base64
import os

import requests

MODELS = ["google/gemini-nano-banana-2.1", "google/gemini-3.1-flash-image"]
PROMPT = "A poster of a lighthouse on a rocky coast at dusk, clean flat style"

for model in MODELS:
    response = requests.post(
        "https://llmtr.com/v1/images/generations",
        headers={"Authorization": "Bearer " + os.environ["LLMTR_API_KEY"]},
        json={"model": model, "prompt": PROMPT, "size": "1K", "aspect_ratio": "3:4"},
        timeout=120,
    )
    response.raise_for_status()
    name = model.split("/")[1] + ".png"
    with open(name, "wb") as image_file:
        image_file.write(base64.b64decode(response.json()["data"][0]["b64_json"]))
    print(model, "->", name)

Switching and getting started on LLMTR

Moving from Nano Banana 2 to 2.1 means changing the model value from google/gemini-3.1-flash-image to google/gemini-nano-banana-2.1; the request format, reference fields and response stay the same. If a request uses the 512 size, set it to 1K. Moving back is the same change.

To start, open an account at llmtr.com, top up prepaid credit and create a key under Dashboard > API Keys. The same key works for every model in the catalog, and each request is deducted from credit at the catalog price in effect when it is processed.

• Google, Gemini Nano Banana 2.1 model page · 2026-10-10

Compare Nano Banana 2.1 and Nano Banana 2 with the same prompt

Create an LLMTR key and send the same prompt to google/gemini-nano-banana-2.1 and google/gemini-3.1-flash-image.

1. Account and key. Open an account at llmtr.com, top up credit and create a key under Dashboard > API Keys.
2. Same request. Send a separate request for each model to https://llmtr.com/v1/images/generations with the same prompt, size and aspect_ratio.
3. Compare. Review the images side by side and compare the cost of the two requests on the usage page in the Dashboard.

Frequently asked questions

Is Nano Banana 2.1 cheaper than Nano Banana 2?

For image output, yes: a 1K image is $0.0336 vs $0.067 and a 4K image $0.113 vs $0.151. Input ($1.50 vs $0.50) and thinking ($7.50 vs $3) cost more on Nano Banana 2.1, so the total depends on the length of your prompts and references.

Does the Nano Banana 2 id keep working on LLMTR?

Yes. google/gemini-3.1-flash-image can still be called. To move to Nano Banana 2.1, change only the model value to google/gemini-nano-banana-2.1.

Can Nano Banana 2.1 generate 512 images?

Google publishes 1K, 2K and 4K sizes for Nano Banana 2.1. The 512 size exists on Nano Banana 2 at $0.045 per image; on 2.1 a 1K image is equivalent to $0.0336.

Related posts

• Gemini Nano Banana 2.1 API: 1K, 2K and 4K image pricing and how to call it
• Editing with reference images, multi-reference and try-on in Nano Banana 2.1

• Gemini Nano Banana 2.1 model
• Gemini 3.1 Flash Image (Nano Banana 2) model
• Google image generation documentation