How to allow for multiple images to be created from a single API call using gemini?

Question

I am currently have a tool that uses Gemini-2.5-flash-image to generate images. However, I am only getting 1 image to be generated. I want the model to create multiple images for example 3 images at once. What is the ideal way to set it up ?

Is it only possible by looping and calling the model multiple times to get more images ? but that seems inefficient to make multiple calls. Seems like this should be possible within a single call.

It also says in the docs (https://docs.cloud.google.com/vertex-ai/generative-ai/docs/models/gemini/2-5-flash-image): “Maximum number of output images per prompt: 10”

Below is the code Ive setup.

Things ive tried:

Changing candidate_count value doesn’t create more images.
I also changed the prompt and asked it to directly created 3 images but still only creates 1 image.

generation_config = GenerateContentConfig(
            response_modalities=[Modality.IMAGE],
            candidate_count=1,  # Only one candidate is supported.
            safety_settings=[
                SafetySetting(
                    category=HarmCategory.HARM_CATEGORY_DANGEROUS_CONTENT,
                    threshold=HarmBlockThreshold.BLOCK_ONLY_HIGH,
                ),
            ],
        )

response = client.models.generate_content(
                    model="gemini-2.5-flash-image",
                    contents=f"Generate an image of: {prompt}",
                    config=generation_config,
                )

Yes I tried that and asked to generate multiple images but it still only provides 1 image. — Coder1934
– Coder1934, Commented Oct 30 at 7:54

Gang Chen · Accepted Answer · 2025-10-30 16:25:52Z

Multiple generation result ≠ candidate_count.

You can generate multiple images results from gemini-2.5-flash-image by doing the following:

Don't set candidate_count
Set response modality have both "TEXT" and "IMAGE"
Prompt message to generate multiple images.

Here is sample snippet to try (works for me):

import base64
import os
from google import genai
from google.genai import types

def generate_and_save_images():
  # Initialize the client. Ensure GOOGLE_CLOUD_API_KEY is set in your environment.
  client = genai.Client(
      vertexai=True,
      api_key=os.environ.get("GOOGLE_CLOUD_API_KEY"),
  )

  model = "gemini-2.5-flash-image"
  prompt_text = "Generate three distinct images: A futuristic city skyline at sunset, a fluffy white cat on a red couch, and a tranquil forest stream."

  contents = [
    types.Content(
      role="user",
      parts=[
        types.Part.from_text(text=prompt_text)
      ]
    ),
  ]

  # Define safety settings
  safety_settings = [
    types.SafetySetting(
      category="HARM_CATEGORY_HATE_SPEECH",
      threshold="OFF"
    ),
    types.SafetySetting(
      category="HARM_CATEGORY_DANGEROUS_CONTENT",
      threshold="OFF"
    ),
    types.SafetySetting(
      category="HARM_CATEGORY_SEXUALLY_EXPLICIT",
      threshold="OFF"
    ),
    types.SafetySetting(
      category="HARM_CATEGORY_HARASSMENT",
      threshold="OFF"
    )
  ]

  # Configure generation parameters. safety_settings is part of GenerateContentConfig.
  generate_content_config = types.GenerateContentConfig(
    temperature = 1,
    top_p = 0.95,
    max_output_tokens = 32768,
    # candidate_count = 1,
    response_modalities = ["TEXT", "IMAGE"],
    safety_settings = safety_settings,
  )

  print(f"Requesting '{generate_content_config.candidate_count}' images from '{model}'...")
  try:
    # Use the non-streaming generate_content method.
    # safety_settings is NOT passed as a separate argument here.
    response = client.models.generate_content(
      model = model,
      contents = contents,
      config = generate_content_config,
    )
except Exception as e:
    print(f"An error occurred during generate_content: {e}")

# Run the function
generate_and_save_images()

Collectives™ on Stack Overflow

How to allow for multiple images to be created from a single API call using gemini?

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related