Exact 9:16 Shorts thumbnail with gpt-image-2.5: use 1152x2048
gpt-image-2.5 on Sume takes custom sizes in multiples of 16. 1152x2048 is exactly 9:16 and above the pixel minimum, while 576x1024 is below it.
Ask gpt-image-2.5 for 1152x2048 for a smaller exact 9:16 file, or 2160x3840 to match YouTube's recommended Shorts size. Both sides must be multiples of 16, which both pairs are. The smaller 576x1024 is also exactly 9:16 but has 589,824 pixels, below the 655,360 pixel minimum for custom sizes.
What YouTube asks
YouTube's help page asks for a 9:16 aspect ratio for Shorts thumbnails, with 2160x3840 recommended (YouTube Help). The 07-24 YouTube blog describes thumbnail generation in Studio for long-form videos (YouTube Blog), so for a Short you make the image yourself and upload it from a computer.
The figures come from the size rules in the Sume image docs.
| Size | Multiple of 16 | Pixels | Ratio | Verdict |
|---|---|---|---|---|
| 2160x3840 | Yes | 8,294,400 | 9:16 | Valid, at the 8,294,400 maximum |
| 1152x2048 | Yes | 2,359,296 | 9:16 | Valid |
| 576x1024 | Yes | 589,824 | 9:16 | Below the 655,360 minimum |
The request
Send image_size with the prompt (not size, which is a resolution tier). Describe the layout, since the thumbnail needs room for the subject and text.
{
"model": "openai/gpt-image-2.5",
"prompt": "Vertical poster for episode 3: a lighthouse at dusk, big empty sky at top",
"image_size": "1152x2048"
}Size versus the recommendation
1152x2048 is smaller than YouTube's recommended 2160x3840, but the stated minimum height is 640 pixels, which 2048 clears. 2160x3840 sits exactly on the model's maximum pixel count, and a bigger image costs more output tokens, so use it for the final and the smaller size for drafts. Check the image docs for current pricing.
Upload
Save the image as PNG or JPG and add it in YouTube Studio on a computer. Keep the series look consistent by repeating the style words in each prompt.
Sources
Related posts
More in Media tools
- Facebook feed video ad minimum is 120 px; Sume's smallest is 256
Meta's Facebook feed video spec lists 120 px minimum sides. Sume Timeline renders even sizes from 256 to 2160, so the floor never blocks an upload.
- Two photos in, one 10-second Omni transition out: $1.875 at 1080p
Gemini Omni first and last frame turns two stills into a transition. On Sume, image_url plus end_image_url for 10 seconds at 1080p costs $1.875.
- Black first frame in a Short? Fade-in and your thumbnail frame pick
output.fade_in_seconds in Timeline 1.0 starts the video from black, so frame 0 is a poor thumbnail. Pick a frame after the fade with Sume video frames.
- Freeze the last 1.5 seconds of a Short for an end card
Hold the final beat of a Short as a still: pull the frame at duration minus 1.5 s with Sume video frames, then lay it as a still slot in Timeline 1.0.
Written by Sume