gemini-3.1-flash-image API

Use the exact model ID gemini-3.1-flash-image. The model field must match this page exactly.

Sample outputgemini-3.1-flash-imagePOST /v1beta/models/gemini-3.1-flash-image:generateContent

Parameters: size=2Kratio=16:9

Prompt: The cold room of a flower shop, steel shelves and buckets of cut stems. A figure in an acid-green vinyl coat has just pulled one bucket toward the door. Along that whole shelf the flowers have begun turning to blown glass, petal by petal, the transformation running outward from the bucket like frost: stems still green at the base, heads already clear and ringing, throwing hard coloured caustics across the steel. One hard pink light through the open cooler door, cold blue shade behind. Wide low camera, macro-sharp on the glass edges, real weight, film grain, no text.

Request that produced this sample

{
  "contents": [
    {
      "role": "user",
      "parts": [
        {
          "text": "The cold room of a flower shop, steel shelves and buckets of cut stems. A figure in an acid-green vinyl coat has just pulled one bucket toward the door. Along that whole shelf the flowers have begun turning to blown glass, petal by petal, the transformation running outward from the bucket like frost: stems still green at the base, heads already clear and ringing, throwing hard coloured caustics across the steel. One hard pink light through the open cooler door, cold blue shade behind. Wide low camera, macro-sharp on the glass edges, real weight, film grain, no text."
        }
      ]
    }
  ],
  "generationConfig": {
    "responseModalities": [
      "IMAGE"
    ],
    "imageConfig": {
      "aspectRatio": "16:9",
      "imageSize": "2K"
    }
  }
}

Authentication and endpoint

Use a Modelflare API key with https://modelflare.dev. The gateway entry point is POST /v1/models/{model}:generateContent.

HTTP

Authorization: Bearer YOUR_MODELFLARE_API_KEY

Generate an image

bash

curl "https://modelflare.dev/v1/models/gemini-3.1-flash-image:generateContent" \
  -H "Authorization: Bearer YOUR_MODELFLARE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "contents": [{"role": "user", "parts": [{"text": "A white cat runs through a neon-lit street in the rain"}]}],
    "generationConfig": {"responseModalities": ["TEXT", "IMAGE"], "imageConfig": {"aspectRatio": "16:9", "imageSize": "1K"}}
  }'

Request notes

Gemini image models use native contents and generationConfig; edits send the text instruction and inline_data in the same parts[].

Request parameters

Gemini image models use native Gemini generateContent. Put text and reference images in the same contents[].parts[]; images use inline_data. Do not send these models to /v1/images/generations, /v1/images/edits, or OpenAI /v1/responses. The imageSize enum is model-specific and shown below.

Field	Required	Notes
contents[].parts[].text	Yes	Text prompt or edit instruction
contents[].parts[].inline_data	No	Base64 image data and MIME type
contents[].parts[].inline_data.mime_type	No	image/jpeg / image/png / image/webp
contents[].parts[].inline_data.data	No	Base64-encoded reference image content
generationConfig.responseModalities	Yes	["IMAGE"] or ["TEXT", "IMAGE"]
generationConfig.imageConfig.aspectRatio	No	1:1 / 4:3 / 3:4 / 16:9 / 9:16
generationConfig.imageConfig.imageSize	No	512 / 1K / 2K / 4K

Response and input limits

Flash Image supports 512, 1K, 2K, and 4K with 1:1, 4:3, 3:4, 16:9, and 9:16. The official documentation does not publish a reference-image count cap, so the gateway adds no local cap.

Image editing example

Put the edit instruction and the reference image in the same contents[].parts[]. inline_data.data is the Base64 image content; do not send Gemini image models to the OpenAI Images or Responses endpoints.

bash

curl "https://modelflare.dev/v1/models/gemini-3.1-flash-image:generateContent" \
  -H "Authorization: Bearer YOUR_MODELFLARE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "contents": [{"role": "user", "parts": [
      {"text": "Keep the subject and lighting, but replace the background with a misty pine forest."},
      {"inline_data": {"mime_type": "image/jpeg", "data": "BASE64_IMAGE_DATA"}}
    ]}],
    "generationConfig": {"responseModalities": ["TEXT", "IMAGE"], "imageConfig": {"aspectRatio": "1:1", "imageSize": "1K"}}
  }'

Response shape

Read candidates[].content.parts[]; do not assume every part is an image because the response may contain text and image parts together.

Field	Notes
candidates[].content.parts[].inline_data	Generated image MIME type and Base64 data
candidates[].content.parts[].text	Text caption returned when TEXT was requested
candidates[].finishReason	Completion or safety reason
usageMetadata	Input/output token usage when returned by the service

Modelflare preserves the native Gemini response shape for gemini-3.1-flash-image and persists usable image parts.

Errors and failed tasks

Field errors, size/ratio limits, and safety refusals from the native Gemini request remain Gemini-compatible error responses. The gateway does not rewrite them as OpenAI Images responses.

Supported scope

This page covers Modelflare model generation and editing calls. Search or map enrichment, voice management, and element-library management are resource APIs outside this public model-call surface and are not required request fields.

Pricing and billing

This model uses token-based billing. See the model pricing page for the current customer price, reference list price, and complete specification.

Note

This page documents Modelflare’s public interface; the fields and limits shown here are the current contract.