/v1beta/openai/images/generations OpenAI “compatible” endpoint returns no token usage, so image generations through the OpenAI compatibility surface cannot be metered.
curl -s -X POST 'https://generativelanguage.googleapis.com/v1beta/openai/images/generations' \
-H "Authorization: Bearer $GEMINI_API_KEY" \
-H 'Content-Type: application/json' \
-d '{
"model": "gemini-2.5-flash-image",
"prompt": "a small red circle on white",
"response_format": "b64_json",
"n": 1
}'
Returns 200 with only:
{
"data": [{"b64_json": "iVBORw0KGgoAAAANSUhEUg..."}],
"model": "gemini-2.5-flash-image"
}
There is no usage block at all. Without it there is no way to attribute cost per request on this endpoint. This prevents usage of this endpoint behind an enterprise gateway.