gemini-2.5-flash-image API

What is gemini-2.5-flash-image?

Use the public identifier gapgpt-gemini-25-flash-image-a96524373c for gemini-2.5-flash-image in ServerNet. The gateway uses this identifier to select the configured route; the developer is Google.

For gemini-2.5-flash-image, validate the request against the gateway’s text input contract. Features of a native Gemini API may require a different integration.

Where to use gemini-2.5-flash-image

• Evaluate gemini-2.5-flash-image on representative visual tasks before selecting it.
• Check dimensions, input image requirements and output format for gemini-2.5-flash-image.
• Keep the original prompt and task settings so visual results can be reproduced.

Build your integration

Create a project key with ai:native and send this page’s example to /v1/inference. Set the model field to gapgpt-gemini-25-flash-image-a96524373c. Store X-Request-Id and X-ServerNet-Charge-Irt with the result, and compare the final amount with your account’s usage receipt. Reuse the same Idempotency-Key only for an identical retry. If the result contains a file link, download and save the file when you receive it. Keeping the response for replay for 24 hours does not extend the supplier’s file-link lifetime.

Inputs and service limits

Keep the input and request options within this route’s documented limits. Check GET /v1/models before sending and use the matching key permission. The reserved amount bounds admission; the usage receipt records the final settled amount.

A bounded request cost is reserved before sending. The verified tariff or reported provider cost is settled using the exchange rate and sale markup. Media is not billed as chat tokens. Use Data URLs for valid images, PCM WAV audio, or H.264 MP4 video on routes that accept video input.

Accepted ServerNet API parameters

Gateway request contract: 2026-10-11. These fields and defaults are enforced by /v1/inference; availability is shown above. Text byte limits are measured as UTF-8 bytes.

Connect to gemini-2.5-flash-image API

Enable ai:native on your key and ai:models:read to check availability. Set SERVERNET_API_KEY in your server environment. Change the idempotency key for each new request; keep it unchanged only for an identical retry.

curl https://servernet.cloud/v1/inference \
  -H "Authorization: Bearer $SERVERNET_API_KEY" \
  -H 'Content-Type: application/json' \
  -H 'Idempotency-Key: CHANGE_FOR_EACH_NEW_REQUEST' \
  -d '{
    "model": "gapgpt-gemini-25-flash-image-a96524373c",
    "n": 1,
    "size": "1024x1024",
    "prompt": "Hello from ServerNet"
}'

Use the same Idempotency-Key when retrying an identical request. Successful responses are replayed without a second charge for 24 hours. Do not automatically retry when x-should-retry:false is present. The key is shown once. Store it in a server environment variable. Keep it out of browser code, GitHub and logs.

Compare models for the same task