Is image-to-video cheaper than text-to-video? Same rate

Image-to-video is not cheaper than text-to-video on Sume. On Gemini Omni Flash 1.1 both bill the same price per output second, so a 5-second clip at 720p is $0.625 either way. If you also generate the still, add that cost: an Ideogram 4.5 still at medium is $0.075, which makes the two-step route $0.70.

The video router docs describe one catalog ID for Omni, routed by the shape of the request: text-to-video from a prompt, image-to-video from image_url with an optional end_image_url, and the same 3 to 10 seconds envelope for both. fal lists $0.10 per second at 720p (read 2026-10-07); Sume bills list times 1.25.

Side-by-side

The table prices each route for a 5-second clip.

When the extra step pays for itself

The still adds between $0.0375 and $0.275, or 6 to 44 percent of the clip. That is a small price for a first frame you have already approved.

A rejected clip costs $0.625 at 720p, so one avoided retake pays for several stills. If a prompt works from text alone about as often as a prompt plus a still, text-to-video is simpler. If the product has to match a photo, start from the photo.

Use the supplied photo route when the product already exists as a photo. Sume takes a public HTTPS image URL, so a photo you host is free of any generation cost.

• Same rate: no discount for starting from an image.
• Resolution moves the price, not the route.
• The still is a separate job with its own reserve.

Two more Omni cases

A start and end frame is still image-to-video with an optional end_image_url, and it bills the same per second. An 8-second 720p clip between two supplied frames is $1.00, and 8 seconds at 1080p is $1.50.

Reference-to-video takes up to 10 reference images and up to 3 reference videos of at most 3 seconds each, with the same envelope and the same per-second billing. More references do not raise the rate in the docs; they are input, and the price comes from output seconds and resolution.

Video-to-video edit is also billed per output second, with 720p as its default resolution. The prompt gives the edit instructions, and there is no duration field.

Why the same rate

The Omni catalog entry bills per output second, as a function of the resolution. The input is a prompt, an image, a set of references or a source video, and none of those is a price line on its own. For budgeting, count seconds and pick the resolution; the route is a quality decision, not a cost one.

Budget line

For 30 products, one 5-second 720p clip each from a supplied photo is $18.75. From a medium Ideogram still it is $21.00, which is $2.25 more. See Wan versus Omni for the other model, and per-image list prices for stills.