Referring to multiple images in vision API

Sending a text message with an ID before each image seems to be working fine:

Main text message:

Analyze the attached images and select the best one for a finance site.

Return the results in JSON format using the following interface:
{
  images: {
    // The id of the analyzed image.
    "id": number;
    // Set this to true if it's the best image.
    "best_image": boolean;
    // One sentence feedback on why you chose the image.
    "feedback": string;
  }[];
}

Here is how i build the rest of the content in PHP:

foreach ($images as $id => $url) {
  $content[] = ['type' => 'text', 'text' => "ID for the next image: $id"];
  $content[] = ['type' => 'image_url', 'image_url' => ['url' => $url]];
}