Gemini Flash Experimental Image Generation and Editing APIs#

In March 2025, Google introduced native image generation and editing capabilities in the Gemini 2.0 Flash Experimental model. You can now generate and iteratively edit images via a single REST endpoint (Experiment with Gemini 2.0 Flash native image generation, Generate images | Gemini API | Google AI for Developers).

How to use Latest Gemini 2.0 Native Image Generation with API? (9 min) (How to use Latest Gemini 2.0 Native Image Generation with API?)

Simple image generation#

To generate a basic image, send a POST request to the generateContent method:

curl "https://generativelanguage.googleapis.com/v1beta/models/gemini-2.0-flash-exp-image-generation:generateContent?key=$GEMINI_API_KEY" \
  -H "Content-Type: application/json" \
  -X POST \
  -d '{
    "contents": [{ "parts": [{ "text": "A serene landscape of rolling hills at sunrise, digital art" }] }],
    "generationConfig": { "responseModalities": ["TEXT", "IMAGE"] }
  }' | jq -r '.candidates[].content.parts[] | select(.inlineData) | .inlineData.data' | base64 --decode > image.png

Replace $GEMINI_API_KEY with your key. (Gemini API | Google AI for Developers)

Generation options#

You can tweak the output with these generationConfig parameters:

  • responseModalities: Modalities to return (TEXT, IMAGE).
  • temperature (0.0–2.0): Controls randomness (default 1.0).
  • topP (0.0–1.0): Nucleus sampling threshold.
  • topK: Token selection cutoff.
  • maxOutputTokens: Max tokens for text parts.
  • stopSequences: Sequences to end generation.
  • seed: For reproducibility.
curl "https://generativelanguage.googleapis.com/v1beta/models/gemini-2.0-flash-exp-image-generation:generateContent?key=$GEMINI_API_KEY" \
  -H "Content-Type: application/json" \
  -X POST \
  -d '{
    "contents": [{ "parts": [{ "text": "A futuristic city skyline at dusk, neon lights" }] }],
    "generationConfig": {
      "responseModalities": ["TEXT", "IMAGE"],
      "temperature": 0.7,
      "topP": 0.9,
      "maxOutputTokens": 1024
    }
  }' | jq -r '.candidates[].content.parts[] | select(.inlineData) | .inlineData.data' | base64 --decode > image.png

Image Generation Docs

Simple image editing#

To edit an existing image, include it in the contents as inlineData (base64-encoded):

curl "https://generativelanguage.googleapis.com/v1beta/models/gemini-2.0-flash-exp-image-generation:generateContent?key=$GEMINI_API_KEY" \
    -H 'Content-Type: application/json' \
    -d '{
      "contents": [{
        "parts":[
            {"text": "Replace the background with a starry night sky"},
            {"inline_data": {"mime_type":"image/jpeg", "data": "'$(base64 -w 0 cat.jpg)'"}}
        ]
      }],
      "generationConfig": {"responseModalities": ["TEXT", "IMAGE"]}
    }' | jq -r '.candidates[].content.parts[] | select(.inlineData) | .inlineData.data' | base64 --decode > image.png

Image Editing Docs

Editing options#

Editing requests support:

  • inlineData: Embed raw image bytes.
  • fileData: Reference public URLs.
  • All generationConfig options listed above.
  • safetySettings: Per-request safety rules.
  • Multi-turn edits by repeating contents in conversation history.

Costs and optimization#

Gemini 2.0 Flash Experimental uses token-based billing:

To optimize:

  • Use smaller image sizes by setting responseMimeType.
  • Cache or reuse prompts with cachedContent.
  • Lower candidateCount or temperature for fewer tokens.

OpenAI gpt-image-1 Model for Image Generation and Editing#

OpenAI’s GPT Image 1 (gpt-image-1) is a state-of-the-art multimodal model released on April 23, 2025, for high-fidelity image creation and editing.

OpenAI’s New GPT Image Model API in 5 Minutes (5 min)

Simple image generation#

Use the Image Generations endpoint:

curl 'https://api.openai.com/v1/images/generations' \
  -H 'Content-Type: application/json' \
  -H "Authorization: Bearer $OPENAI_API_KEY" \
  -d '{
    "model": "gpt-image-1",
    "prompt": "A whimsical illustration of a cat playing chess",
    "n": 1,
    "size": "1024x1024"
  }' > image.png

(Generate Image | OpenAI API - Postman)

Generation options#

Adjust these JSON parameters:

  • model: gpt-image-1 (default).
  • prompt: Text description.
  • n: Number of images.
  • size: 256x256, 512x512, or 1024x1024.
  • response_format: "url" (default) or "b64_json".
{
  "model": "gpt-image-1",
  "prompt": "...",
  "n": 2,
  "size": "512x512",
  "response_format": "b64_json"
}

Simple image editing#

Use the Edits endpoint with an image and a mask:

curl https://api.openai.com/v1/images/edits \
  -H 'Content-Type: application/json' \
  -H "Authorization: Bearer $OPENAI_API_KEY" \
  -d '{
    "model": "gpt-image-1",
    "image": "data:image/png;base64,<BASE64_IMAGE>",
    "mask": "data:image/png;base64,<BASE64_MASK>",
    "prompt": "Add a rainbow in the sky above the mountains",
    "n": 1,
    "size": "1024x1024"
  }'

(curl - What’s the correct URL to test OpenAI API? - Stack Overflow)

Editing options#

Editing requests accept:

  • image: Original image (base64 or URL).
  • mask: PNG mask for inpainting.
  • prompt: Instruction for the edit.
  • n, size, response_format as above.
  • Optional user field for attribution.

Costs and optimization#

GPT Image 1 pricing (per 1M tokens): text input $5, image input $10, image output $40. Rough per-image costs:

To optimize:

  • Choose smaller sizes (256x256).
  • Generate fewer images (n:1).
  • Use response_format:"url" to reduce payload.
  • Cache frequent prompts or images.
Backward Vision Models Llm Speech Forward