Gemini Image Generation

Generate high-quality images from text prompts using Google's Gemini 3 Pro model via fal.ai.

Prerequisites

• FAL_KEY environment variable must be set (typically in ~/.zshrc)

API Endpoint

POST https://fal.run/fal-ai/gemini-3-pro-image-preview

Parameters

Required

• prompt (string): The text description of the image to generate

Optional

Usage

cURL

curl --request POST \
  --url https://fal.run/fal-ai/gemini-3-pro-image-preview \
  --header "Authorization: Key $FAL_KEY" \
  --header "Content-Type: application/json" \
  --data '{
    "prompt": "A serene mountain landscape at sunset with golden light",
    "num_images": 1,
    "aspect_ratio": "16:9",
    "resolution": "2K",
    "output_format": "png"
  }'

Python

import fal_client

result = fal_client.subscribe(
    "fal-ai/gemini-3-pro-image-preview",
    arguments={
        "prompt": "A serene mountain landscape at sunset with golden light",
        "num_images": 1,
        "aspect_ratio": "16:9",
        "resolution": "2K"
    }
)

# Access the generated image URL
image_url = result["images"][0]["url"]
print(f"Generated image: {image_url}")

JavaScript

import { fal } from "@fal-ai/client";

const result = await fal.subscribe("fal-ai/gemini-3-pro-image-preview", {
  input: {
    prompt: "A serene mountain landscape at sunset with golden light",
    num_images: 1,
    aspect_ratio: "16:9",
    resolution: "2K"
  }
});

console.log("Generated image:", result.images[0].url);

Response Format

{
  "images": [
    {
      "file_name": "generated_image.png",
      "content_type": "image/png",
      "url": "https://storage.googleapis.com/..."
    }
  ],
  "description": "A description of the generated image"
}

Examples

1. Simple image generation:

   • Prompt: "Generate an image of a futuristic city at night"

2. Specific aspect ratio:

   • Prompt: "Create a portrait-oriented image of a forest path" with aspect_ratio: "9:16"

3. High resolution:

   • Prompt: "Generate a detailed 4K image of a coral reef" with resolution: "4K"

Tips

• Be specific in your prompts for better results
• Include lighting, mood, and style descriptors
• Use appropriate aspect ratios for your use case (16:9 for landscapes, 9:16 for portraits)
• Higher resolution takes longer to generate

Error Handling