Amazon A+ hotspot image module: make the 1464x600 art first

Amazon's PremiumHotspotImageModule is a full-width image with clickable markers that reveal a headline and text for each point of interest. The desktop image must be at least 1464 by 600 pixels. Generate it once at its final size, 1472 by 608 on openai/gpt-image-2.5, and place hotspots against that exact image.

Amazon's description is from its A+ Premium modules page (the snapshot did not include the hotspot coordinate fields, so check Amazon's full spec for them); Sume's size rules are from the image generation docs, read 2026-10-01.

What is the hotspot module for?

Amazon lists two uses: simplify and structure feature callouts for complex, feature-heavy products, and answer questions like what a button does or what a material feels like. The image is the stage, so it must show every point you plan to mark.

Why generate at the final size?

Markers are placed on the image, so if you resize or crop later, each marker's position can drift. Producing the art at the size you will upload removes that step. The size is the one in the full-background guide linked below.

{
  "model": "openai/gpt-image-2.5",
  "prompt": "Top-down view of an espresso machine, controls clearly visible, calm margins",
  "image_size": "1472x608"
}

What goes in the image?

Leave clear, uncluttered areas around each feature you will mark, so a marker does not cover the detail it points at. Ask for the feature in the prompt, for example the button, the hinge or the fabric, and review the result before you pin markers.

Does this size rule apply to all models?

No. The multiples-of-16 rule is documented for the two ChatGPT Image 2.5 ids. For other models use the normalized tier and ratio and read supported_parameters. Related: Premium A+ full background image.