Guide to calling gemini-2.5-flash-image (Nano Banana)

This document demonstrates how to use Google Gemini's image model, gemini-2.5-flash-image-preview, via cometapi for image generation. It covers two common methods:

Gemini's official generateContent API for text-to-image generation

Gemini's official generateContent API for image-to-image generation (both input and output are Base64)

Important Notes:

Replace sk-xxxx in the examples with your cometapi key. For security, do not expose your key in client-side code or public repositories.

The Authorization header for cometapi typically uses the key value directly, e.g., Authorization: sk-xxxx.

The returned image is usually provided as Base64-encoded inline_data. You will need to decode it on the client side and save it as a file.

Basic Information:

Base URL: https://api.cometapi.com

Model Name: gemini-2.5-flash-image-preview / gemini-2.5-flash-image

I. Gemini's Official `generateContent` for Text-to-Image

Use Gemini's official generateContent endpoint for text-to-image generation. Place the text prompt in contents.parts[].text.

Example (Windows shell, using ^ for line continuation):

Response Highlights:

The image data is typically found in response.candidates[0].content.parts, which can contain:

Text description:

{ "text": "..." }

Image data:

{ "inline_data": { "mime_type": "image/png", "data": "<base64>" } }

Decode the data field (the Base64 string) and save it as a file with the corresponding extension.

II. Gemini's Official `generateContent` for Image-to-Image (Base64 I/O)

This endpoint supports "image-to-image" generation: upload an input image (as Base64) and receive a modified new image (also in Base64 format).

Example:

Description:

First, convert your source image file into a Base64 string and place it in inline_data.data. Do not include prefixes like data:image/jpeg;base64,.

The output is also located in candidates[0].content.parts and includes:

An optional text part (description or prompt).

The image part as inline_data (where data is the Base64 of the output image).

For multiple images, you can append them directly, for example:

III. Official Gemini: Image Generation from Multiple Images (Base64 Input/Output)

This endpoint supports "multi-image to image" generation: upload multiple input images (Base64) and it will return a new, modified image (also in Base64 format).

Method 1: Combine multiple images into a single collage, as shown in the example below

Example input description:

A model is posing and leaning against a pink bmw. She is wearing the following items, the scene is against a light grey background. The green alien is a keychain and it's attached to the pink handbag. The model also has a pink parrot on her shoulder. There is a pug sitting next to her wearing a pink collar and gold headphones.

Returned Base64 converted back to an image:

Example:

Notes:

First, convert your source image file to a Base64 string and insert it into inline_data.data (do not include prefixes like data:image/jpeg;base64,).

The output is also located in candidates[0].content.parts and contains:

An optional text part (description or prompt)

An image part inline_data (where data is the Base64 of the output image)

Method 2: Pass multiple images via Base64 parameters (up to three images)

Example:

API Input Parameters:

Example of returned image:

How to Extract and Save the Base64 Image from the Response

Using the Gemini-style response as an example, the pseudo-structure is as follows (for illustration purposes):

{
  "candidates": [
    {
      "content": {
        "parts": [
          { "text": "..." },
          {
            "inlineData": {
              "mimeType": "image/png",
              "data": "<base64-string>"
            }
          }
        ]
      }
    }
  ]
}

Extract the inlineData.data string and choose the file extension based on its mimeType (e.g., .png, .jpg, .webp).

In your client application, decode the Base64 string and save it as a binary file.

Python Example:

Node.js Example:

FAQ & Suggestions

Authorization Header Format

Use Authorization: sk-xxxx. Please refer to the official cometapi documentation for confirmation.

Prompt Optimization

Specifying style keywords (e.g., "cyberpunk, film grain, low contrast, high saturation"), aspect ratio (square/landscape/portrait), subject, background, lighting, and level of detail can help improve the results.

Base64 Note

Do not include a prefix like data:image/png;base64, in the data field; only include the pure Base64 data string.

Troubleshooting

4xx errors usually indicate issues with request parameters or authentication (check your key, model name, JSON format). 5xx errors are server-side problems (you can retry later or contact support).

The above covers the methods and key points for using gemini-2.5-flash-image-preview via cometapi. Choose the API style that suits your needs and implement the Base64 image decoding and display on the client side.

🍌 Flash 2.5 Image Updates

a. Flexible Aspect Ratios

Now supports multiple aspect ratio settings for easy content creation across different devices. All resolutions consume 1,290 tokens by default.

Supported aspect ratios:
1:1, 3:2, 2:3, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9, 21:9

Reference examples:

https://ai.google.dev/gemini-api/docs/image-generation#aspect_ratios

https://cloud.google.com/vertex-ai/generative-ai/docs/multimodal/image-generation#googlegenaisdk_imggen_mmflash_with_txt-drest

https://github.com/GoogleCloudPlatform/generative-ai/blob/main/gemini/getting-started/intro_gemini_2_5_image_gen.ipynb

b. Model Name Update

All new features will be available on the new model ID: gemini-2.5-flash-image. The previous gemini-2.5-flash-image-preview will be deprecated.

⚠️ Migration required by October 31, 2025

c. Force Image Output

To address the frequent issue of text-only outputs, you can now set "responseModalities" to ["IMAGE"] only in API requests. This ensures image generation without text-only responses.
Official API documentation for reference:

Image Generation

Image Understanding (Multimodal)

Image Generation

Guide to calling gemini-2.5-flash-image (Nano Banana)

I. Gemini's Official generateContent for Text-to-Image#

II. Gemini's Official generateContent for Image-to-Image (Base64 I/O)#

III. Official Gemini: Image Generation from Multiple Images (Base64 Input/Output)#

Method 1: Combine multiple images into a single collage, as shown in the example below#

Method 2: Pass multiple images via Base64 parameters (up to three images)#

How to Extract and Save the Base64 Image from the Response#

FAQ & Suggestions#

🍌 Flash 2.5 Image Updates#

a. Flexible Aspect Ratios#

b. Model Name Update#

c. Force Image Output#