Avatar product image adds about 5% per second on every Sume tier
Adding a product_image to a Sume avatar video raises the per-second rate by 1 cent on standard, 1.3 cents on plus, 3 cents on max. The table and a 30 s example.
A product image on an Avatar 1.0 talking video adds about 5% to the per-second rate on all three tiers: $0.01 per second on standard, $0.013 on plus, and $0.03 on max. If you leave out product_image, you pay the no-product rate. Sume avatars are async jobs: you submit a request, get a job id back, and read the finished video later. There is no live video session.
The rates come from AVATAR_VIDEO_USD_MICROS_PER_SECOND in packages/provider-pricing, rate card sume_avatar_video_fixed_tier_rates_2026_07_02.
The rate table
| Tier | No product | With product | Difference | Increase |
|---|---|---|---|---|
| standard | $0.184 | $0.194 | $0.01 | 5.4% |
| plus | $0.245 | $0.258 | $0.013 | 5.3% |
| max | $0.55 | $0.58 | $0.03 | 5.5% |
A 30-second example
At 30 seconds the product surcharge is small compared with the tier choice itself.
| Tier | No product | With product | Extra |
|---|---|---|---|
| standard | $5.52 | $5.82 | $0.30 |
| plus | $7.35 | $7.74 | $0.39 |
| max | $16.50 | $17.40 | $0.90 |
When the surcharge is worth it
Use product_image when the video is about an item and the viewer should see it: a UGC-style ad, a product explainer, a demo hook. Leave it out for a pure talking-head message such as an announcement, a welcome, or a policy update. The surcharge is a fixed amount per second, so a longer video pays it for every second even if the product is only on screen for part of it. If only a short beat needs the product, a multi-scene video_inputs plan or a separate short clip can keep the average cost down, but check the current rules before you split: the current execution supports one resolved avatar for each final video and expects the scene backgrounds to resolve to one shared scene.
The request
Media fields must be fetchable public HTTPS URLs. Sume rejects localhost, private-network URLs, non-HTTPS URLs, and non-image responses before it submits generation.
curl -X POST https://api.sume.com/v1/avatar-1.0/talking-video \
-H "Authorization: Bearer $SUME_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: bottle-ad-001" \
-d '{
"avatar_handle": "product_host",
"script": "This bottle keeps drinks cold for a full day.",
"product_image": "https://example.com/bottle.png",
"quality": "plus",
"aspect_ratio": "9:16"
}'Budgeting a batch with products
Suppose you make a 20-second product clip for each of 25 items. At standard with a product image, that is 25 × 20 seconds × $0.194, which is $97.00. The same batch at plus is $129.00 and at max it is $290.00. Without product images, the same three batches would be $92.00, $122.50 and $275.00.
The tier moves the cost far more than the product image does. Going from standard to max multiplies the with-product rate by about 3.0. Going from no product to product multiplies it by about 1.05. So choose the tier by how much the quality matters for the placement, and use the product image whenever the item is the subject.
A tip for drafts
Review the first frame with an Avatar video preview before the full render. The preview is tier-independent and Sume reuses the stills at the final render, so you can approve the composition, including the product placement, and then pick the final tier with quality on generate-video. Structural changes, such as a different avatar or scene, need a new preview.
The field is product_image, and the Generate avatar video page says to omit it for a productless video. Use a clean, public HTTPS image of the item, ideally on a plain background. Because Sume fetches the URL before it submits the generation, a link that needs a login, points at a private network, or returns a page instead of an image is rejected up front, and you find out before any provider work starts.
If you also want a scene reference, that is a separate field. scene takes either a prompt ({"type":"prompt","prompt":"..."}) or a photo ({"type":"photo","image_url":"..."}). The docs list product_image and scene as independent optional inputs, so you can use either, both, or neither. Check the rate table above for the one you are paying for: the page that lists the per-second rates distinguishes only product from no product.
Sources
Related posts
More in Pricing
- GPT Image 2.5 at 4:5 lists 12-14% below 1:1 at the same quality
On Sume, GPT Image 2.5 at 4:5 (1024x1280) lists 12 to 14 percent below a 1024x1024 square at every quality tier. The five-tier table, with billed prices.
- GPT Image 2.5 quality ladder: price multiplier from low to max on Sume
At 1024x1024 on Sume, GPT Image 2.5 high costs 8.9 times low, xhigh 15.9 times and max 35.7 times. The ladder with billed prices and batch costs.
- How much do six gift guide cover images cost at each quality tier?
Six Ideogram 4.5 cover options on Sume: $0.225 at low, $0.45 at medium, $1.65 at high. Which tier to use for which step.
- How many AI music tracks should I generate for one ad before picking?
Each Music Router track is a flat $0.125, so three takes cost $0.375 and eight cost $1.00. Here is a take budget by how fixed your brief is.
Written by Sume