Ideogram 4.5 YouTube thumbnail: 16:9 headline and the 2 MB limit

Ask ideogram/ideogram-v4.5 for aspect_ratio: "16:9" with the headline in quotes in the prompt, then check the file against YouTube's rules: 16:9 recommended, JPG or PNG, and 2 MB on mobile. Sume lists Ideogram at three quality tiers; the 1K or 2K choice changes pixels, not list price.

What YouTube asks for

The YouTube thumbnail help page recommends 3840x2160 for videos with 16:9, a minimum width of 640 pixels, and JPG or PNG files. The size limit is 2 MB on mobile and 50 MB on desktop. Treat the 2 MB number as your ceiling if you upload from a phone.

Generate the headline

Ideogram is the model in the Sume catalog built around text in images. Use resolution: "2K" for the larger file and quality: "medium" (the default) first. Keep the headline to a few words so it stays readable at small sizes.

curl -X POST https://api.sume.com/v1/images \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"ideogram/ideogram-v4.5","prompt":"YouTube thumbnail, the headline \"FIX IT IN 5 MIN\" in bold white type on the left, a laptop and a toolbox on the right","aspect_ratio":"16:9","resolution":"2K","quality":"medium"}'

Cost of trying variants

The repo's Ideogram list prices are $0.03, $0.06 and $0.22 per image for low, medium and high, not dependent on size. Sume bills list times 1.25, rounded up to the cent. Reading usage.cost on each response shows the exact figure.

Iterating on wording at low, then rendering the winner at medium or high, is cheaper than generating every variant at the top tier.

Check the file before upload

Image responses come back as hosted URLs. Download the file and look at its byte size before uploading: a 2K render can exceed 2 MB in PNG. If it does, convert to JPG locally or upload from desktop, where the limit is 50 MB.

Then view it at about the size of a suggested-video tile. If the headline is not legible there, shorten the words and regenerate.

Editing an existing thumbnail

If you already have a thumbnail and only the words need to change, send it as the first input_references entry. Ideogram accepts up to five references and edits the first one. When you leave out aspect_ratio on an edit, the result keeps the source shape, so a 16:9 thumbnail stays 16:9. That makes a title tweak a one-request job instead of a redesign.

Layout rules that survive small sizes

Thumbnails are viewed small, often with a timestamp badge covering the lower-right corner. Put the headline on the left or top, keep it to three or four words, and give it strong contrast with the background. Ask for the exact text in quotes and check the spelling in the output every time: text rendering is much better than it used to be but a mistyped letter still ships if you do not look.

• Three to four words at most.
• Leave the lower-right corner free of text.
• High contrast between type and background.
• Render once at low to test wording before paying for high.

Shorts use a different shape

The same YouTube page recommends 2160x3840 at 9:16 for Shorts thumbnails, so a Shorts cover is a separate request with aspect_ratio: "9:16" rather than a crop of the wide one. Check the page again before you rely on any number here; YouTube changes these rules, and this post only reflects what it said on the day it was read.