Product spin videos for 100 SKUs on $100: cost by Sume model

Animating 100 product photos as 5-second spins costs between $6.25 and $105.00 on Sume, depending on the video model; Wan 3.0 at 480p is $31.25 and Wan 3.0 at 720p is $62.50. The model choice moves the bill by a factor of 17, so spend the first few dollars on a three-SKU test instead of the whole $100.

None of these models is built for a guaranteed 360-degree turntable. A spin is a prompt outcome ("slow orbit around the product, static studio background"), and some models will hold the pose, wobble the label or invent the back of the pack. The table prices the attempt; the test tells you which model is worth the price.

What each 5-second image-to-video clip costs

Each model below takes a product photo as the first frame (image_url) and returns one clip. Rates are the Video Router catalog rate times Sume's 1.25 margin, per output second; the cost column is 5 seconds. Seedance bills by video tokens, so its 5-second price depends on the aspect ratio; the Mini rows use 1:1 at 480p and 720p, the shape of most catalog photos.

Ratios and limits that decide the shortlist

Catalog photos are usually square or 4:5. Wan 3.0, MiniMax H3, Kling 3.0 and Seedance all list 1:1, so a square photo yields a square clip. Omni accepts only 16:9 and 9:16, which means a square product gets either cropped or padded for a vertical or landscape feed. Grok Imagine Video 1.5 takes no aspect_ratio field at all and needs image_url or first_frame_url, so it is image-to-video only.

Durations also differ: Omni accepts 3 to 10 seconds, Grok, Kling and Seedance 2.0 4 to 15, MiniMax H3 5 to 15, Wan 3.0 2 to 30. A 5-second spin sits inside every range. Kling 3.0 is billed on seconds and audio, not on resolution: 720p and 1080p cost the same, and audio on is the default, so send generate_audio: false for a silent spin and save 33%.

A $100 plan that does not burn it on one model

Take three SKUs that look different, for example glossy, matte and transparent packaging. Run each through Grok, Wan 480p and Wan 720p: 9 clips, $3.00. Add Kling 3.0 with audio off for the same three SKUs, $2.10. For $5.10 in total you see four models on your own products.

Then run all 100 SKUs on the winner. At Wan 480p that is $31.25; allow a 30% retake rate for bad spins and the budget is $40.62. If the winner is Wan 720p, 100 SKUs plus 30% retakes is $81.25, still inside the $100 line. Kling with audio on, at $105.00 for one take of each SKU, does not fit.

Reserve a little for assembly. Join several spins into one catalog reel with a timeline render at $0.10 per output minute, and cut long ones with video trim at $0.02 per job. A 20-SKU reel of 5-second clips is 100 seconds, so two billed minutes, $0.20. For the price spread across other models see the five-second 720p comparison.