Pinterest pin size and video specs: 2:3 images and clips

Pinterest recommends 2:3 or 1000x1500 image Pins and takes 4 s to 15 min video. Make 2:3 images with Sume, and crop 9:16 clips to 2:3 or 4:5.

5 min readSume
All posts

Pinterest's ad specs recommend a 2:3 aspect ratio, or 1000 x 1500 pixels, for image Pins in PNG or JPEG, and take MP4, MOV, or M4V video in H.264 or H.265 from 4 seconds to 15 minutes, with square or vertical 2:3, 4:5, or 9:16 recommended. With the Sume API, send aspect_ratio: "2:3" to an image model that lists it; for video, generate 9:16 or 1:1, then crop to 2:3 or 4:5, because no Sume video model lists those two ratios.

Pinterest's figures are quoted from its Review ad specs help page, read on 2026-09-27. It is an ads page, and it lists standard image and standard width video ads as usable in paid campaigns or for organic impressions. Sume's facts come from the Image API, Video filter, and Video captions docs, the Sume API reference, and Sume's model catalogs.

What are Pinterest's pin image and video specs?

Pinterest warns that “Pins with an aspect ratio greater than 2:3 might get cut off in people's feeds”, and it converts every uploaded image to an 8-bit RGB JPEG.

From Pinterest's Review ad specs page, read 2026-09-27.
FormatFilesShapeLength
Standard image.PNG or .JPEG; 20 MB on desktop, 32 MB in-app2:3 recommended, or 1000 x 1500 pixels—
Standard width video.MP4, .MOV, or .M4V; H.264 or H.265; up to 2 GB“Shorter than 1:2 (width:height), taller than 1.91:1”; 1:1, 2:3, 4:5, or 9:16 recommended4 s to 15 min; 6 to 15 s recommended for ads
Idea ad, full bleedVideo on iOS and Android only; 2 GB in the apps, 100 MB on web1080x1920 (9:16) recommendedUp to 5 minutes

How do I make a 2:3 pin image with the Sume API?

Send POST /v1/images with aspect_ratio: "2:3" to a model whose supported_parameters list it, such as google/nano-banana-pro or google/nano-banana-2, and output_format: "png", since Pinterest advises uploading PNG. A model accepts only the ratios its catalog descriptors list.

The ChatGPT Image models don't list 2:3, but they take custom pixels in image_size, with both edges multiples of 16. 1000x1500 fails that rule; 1024x1536 is the same 2:3 shape at 1,572,864 pixels, inside the docs' 655,360–8,294,400 range. AI image aspect ratios and custom sizes covers image_size.

curl -X POST https://api.sume.com/v1/images \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "google/nano-banana-pro",
    "prompt": "A linen tote bag on a sunlit oak table",
    "aspect_ratio": "2:3",
    "output_format": "png"
  }'

Can Sume generate 2:3 or 4:5 video?

Not directly. No model in Sume's video catalog lists 2:3 or 4:5, and the Video Router's aspect_ratio field has neither. Models that list 9:16 or 1:1, both shapes Pinterest recommends, make them directly. For exact pixels such as 1000x1500, put the clip in a Timeline 1.0 render with even output.width and output.height: its default cover fit fills the frame and crops the overflow, and its sound comes from the render's audio spine, not the clip.

To keep the clip's own audio, crop a 9:16 clip with POST /v1/video-filter instead. It reads a workspace media.sume.com clip of up to 300 seconds, keeps its frame rate and audio, and returns a new MP4; today Sume's compiler encodes it as H.264 (libx264), both on Pinterest's lists, and YouTube Shorts video specs compares the rest of that encode. The crops below keep the full width and the middle of the height; the fractions are our own arithmetic for an exact 9:16 source, and POST /v1/video-filter/check validates one for free. Convert a landscape video to vertical shows the same crop op.

Our arithmetic for an exact 9:16 source, as Video filter crop ops, whose fields are fractions of the source frame; read 2026-09-27.
Pinterest shape`ops` entry for a 9:16 clip1080x1920 becomes
2:3{ "op": "crop", "x": 0, "y": 0.078125, "width": 1, "height": 0.84375 }1080x1620
4:5{ "op": "crop", "x": 0, "y": 0.1484375, "width": 1, "height": 0.703125 }1080x1350

Which clip lengths and caption positions suit Pinterest?

Pinterest's video minimum is 4 seconds, and it recommends 6 to 15 seconds for video ads. Send a duration in that range, and never below 4 on the models whose floor is lower: wan-3.0 accepts 2 seconds and gemini-omni-flash-1.1 accepts 3.

Pinterest's Idea ad specs, which recommend 1080x1920 for full-bleed video, give pixel safe zones: keep text 270 px from the top, 440 px from the bottom, 65 px from the left, and 195 px from the right. On POST /v1/video-captions, design.placement.anchor_ratio sets a caption line's center as a fraction of frame height (0.05–0.95), and design.typography.safe_width_ratio the share of frame width a line may use (0.3–1). By our arithmetic for a 1080x1920 frame, a centered line clears the side zones at a safe_width_ratio of 0.63 or less, with its center well inside 0.15–0.77 of the height. design isn't supported on the punch or tiktok-green styles.

Sources

Related posts

More in Media tools

All Media tools posts

Written by Sume