Facebook Reels Image Ad Specs: 1440x2560, 40 and 55 Characters
Meta lists Facebook Reels image ads as 9:16 at 1440x2560, 30 MB, 600 px minimum, 40/55 text limits. Generate a 9:16 still with the Sume Images API.

What are the Facebook Reels image ad specs?
Meta's page for Awareness image ads on Facebook Reels lists JPG or PNG, 9:16, 1440x2560, 30 MB, a minimum of 600 px on each side, and a 3% tolerance. Primary text is 40 characters and the headline is 55. The safe zone is roughly 14% at the top, 35% at the bottom and 6% at each side.
The page also notes that, because of the small image size, Meta recommends not adding text on the image. I read it on 2026-10-03.
| Field | Meta's value |
|---|---|
| File types | JPG, PNG |
| Aspect ratio and size | 9:16, 1440x2560 |
| Minimum dimensions | 600 px wide and tall |
| Tolerance | 3% |
| Max file size | 30 MB |
| Primary text | 40 characters |
| Headline | 55 characters |
How does it differ from the Instagram Reels image ad?
The Instagram page shares the 1440x2560 size and 30 MB limit but gives a 500 px minimum, a 1% tolerance and 44 characters of primary text, with no headline field in what I read. Two near-identical formats with different numbers is the reason to keep one spec table per placement and not one for all of Meta.
How do I make the still with Sume?
POST /v1/images with model: "sume/auto" and aspect_ratio: "9:16" returns a vertical image. Sume documents custom pixel sizes on one model, gpt-image-2.5, which accepts an image_size with both edges multiples of 16, a maximum edge of 3840 and an aspect of at most 3:1. 1440 and 2560 are both multiples of 16, so 1440x2560 fits those rules. Check the model's catalog entry before relying on it.
Add an Idempotency-Key header to the write, and save the output as PNG or JPEG.
curl https://api.sume.com/v1/images \
-H "Authorization: Bearer $SUME_API_KEY" \
-H "Idempotency-Key: $(uuidgen)" \
-H "Content-Type: application/json" \
-d '{"model":"sume/auto","prompt":"Running shoe on a wet street at dusk, centred","aspect_ratio":"9:16","output_format":"png"}'What should go in the image and what should not?
Meta says not to add text on this image. Keep the subject in the middle band that survives the 14% and 35% margins, and put your offer in the primary text and headline. If you need a still plus motion, compose the still over a clip with Sume's timeline compose endpoint, which takes one still and one video in a stack or overlay at $0.02 per job.
Sources
Related posts
More in Use cases
- Facebook unoriginal content: which edits to a clip still count?
Meta calls borders, captions and speed changes to someone else's clip low-value and deprioritizes them. What counts as original, and what Sume edits change.
- Find the product moment in a Short: a contact sheet with video-frames
For an affiliate or product-tag Short, pull up to 24 frames in one call with video-frames, spot when the product is on screen, then cut to that second.
- Flat-lay garment photo to on-model image with GPT Image 2.5
Turn a flat-lay or hanger photo into an on-model image with openai/gpt-image-2.5 on Sume: reference order, aspect ratio auto, and a quality ladder.
- Flyer to video: turn a promo flyer into a vertical clip
Make a 10-second vertical clip from a flyer image and a short clip of the shop, using Timeline 1.0 and a compose overlay for the offer.
Written by Sume