Meta image animation or your own image-to-video: which to run
Meta Advantage+ can animate a product image; you can animate it yourself on Sume. Compare control, and note Meta's differing label statements on third-party AI.

Run Meta's image animation when you want speed and are content to let Meta choose the motion. Animate the image yourself when the motion matters, for example a lid that must open or a label that must stay readable. With Sume you pick the model, the prompt, the duration and the first frame. Meta's pages are not consistent on whether ads made with third-party AI tools get an AI label, so check the current rule before you launch.
What do Meta's pages say?
The lift figures are Meta's own claims about its features. They say nothing about a clip you animate yourself.
| Topic | Statement |
|---|---|
| Advantage+ creative features | Include image animation, background generation and video expansion |
| Reported lifts | 2-3% conversion lift from background generation on catalog ads; 2% lift on Reels from video expansion |
| Label on generative features | An "AI logo beneath the feature name" |
| When ads are labelled | Help page: ads fully created or materially edited with Meta's generative AI creative features; it says AI info is not applied to ads made with third-party tools. Meta's 2025 transparency post says ads detected as made with third-party AI tools can be labelled. |
When should I animate the image myself?
Do it yourself when one thing must happen on screen. Sume's POST /v1/videos takes your photo as a first_frame in frame_images and a prompt that says what moves. seedance-2.5 accepts 4 to 30 seconds. gemini-omni-flash-1.1 accepts 3 to 10 seconds at 360p up to 4K. Check GET /v1/videos/models for the ratios and durations your chosen model lists.
Make a cheap pass first at 480p, check the product against the photo, and only then render the winner at a higher tier.
{
"model": "seedance-2.5",
"prompt": "The jar lid twists open, steam rises, camera stays still. Keep the label sharp and unchanged.",
"aspect_ratio": "4:5",
"resolution": "480p",
"duration": 5,
"frame_images": [{
"type": "image_url",
"image_url": { "url": "https://cdn.example.com/jar.jpg" },
"frame_type": "first_frame"
}]
}What about the AI label?
Meta's Help page ties AI info to its own generative features and says it is not applied to ads made with third-party tools, while its 2025 transparency post says ads detected as made with third-party AI tools can be labelled. Because the two differ, write copy that does not claim the footage is a real recording, and check the current rules in Meta's ads tools before you launch.
How do I choose?
Use the test that matters most to you, not the one that is easiest to run.
- If you need an exact motion or a set duration, run image-to-video yourself.
- If you want the platform to try several variants for one image, use the platform option.
- If the claim in the ad must match the clip frame by frame, keep the edit under your own control.
- In either case, check Meta's current label rules before the ad is live.
Sources
Related posts
More in Use cases
- Hook, body and CTA from three video models in one 30-second Short
Seedance 2.5 for a 4 s hook, Wan 3.0 for a 20 s body and Omni Flash for a 6 s call to action cost $9.31 in generation, $9.41 with the Timeline join.
- Museum exhibit intros on Wan 3.0: 22 s each, eight for $22.00 at 720p
A 22-second Wan 3.0 clip costs $1.375 at 480p, $2.75 at 720p and $5.50 at 1080p on Sume, the same as two 11-second jobs. One job keeps the shot continuous.
- Nano Banana 2.1 0.5K is a 512 tier; YouTube thumbnails need 640 wide
The Image API names its smallest tier 512, below YouTube's 640-pixel minimum thumbnail width. 91 test thumbnails cost $6.825 at 0.5K, $9.10 at 1K, $13.65 at 2K.
- 120 Nano Banana 2.1 2K hero images: $6.05 at Google, $18.00 on Sume
Google lists Nano Banana 2.1 at $0.0504 per 2K image; Sume lists $0.15. For 120 listing hero images that is $6.05 versus $18.00, plus what the difference buys.
Written by Sume