Veo 3.1 Lite is $0.05 a second: Sume 720p video prices per second

Google's Gemini API pricing page, read on 2026-10-04, lists Veo 3.1 Lite at $0.05 a second at 720p and $0.08 at 1080p, with audio included. If that number is your yardstick, the useful question is what other 720p models cost per second once the price a buyer pays is the number you compare. This post puts Sume's per-second rates next to it. It compares prices only, not quality or availability.

What Sume bills per second at 720p

Sume bills the provider list price times 1.25, and the docs say that applies to every video model. The per-second figures below come from the pricing estimators in Sume's code, run for a 10-second clip and divided by ten. Wan 3.0 and Gemini Omni Flash 1.1 are flat per second; the Seedance models are priced by video tokens, so their per-second rate is the 10-second total divided by ten.

What an 8-second clip costs

Multiply by duration. An 8-second 720p clip is $0.40 at Veo 3.1 Lite's list rate, $1.00 on Wan 3.0 through Sume, and about $4.62 on Seedance 2.5. The gap between Veo Lite and Wan is partly the 1.25 factor and partly a different list price underneath, so the two numbers do not measure the same thing: one is a vendor list price, the other is what lands on a workspace balance.

How to use the comparison

• Compare like with like: audio included or not, resolution, and maximum clip length all differ by model.
• Treat a vendor list price as a floor. Sume adds the 1.25 factor to the list price of whatever model it routes to.
• Read the live price from the models endpoint before you commit a batch; the docs tell you to use GET /v1/videos/models for capabilities and prices.

Why the same resolution is not the same product

Two 720p clips can differ in everything that drives cost. Google's page says Veo 3.1 Lite prices include audio. Sume's Wan 3.0 clips are priced per second of video, and Gemini Omni Flash 1.1 is documented with native synced audio. Kling v3 Pro, by contrast, has a separate audio-on rate, which is a reminder to check whether a headline per-second number already contains sound before you compare it.

Maximum length matters too. Wan 3.0 accepts up to 30 seconds in one clip on Sume, Gemini Omni Flash 1.1 up to 10, and most other catalog models up to 15. A model that stops at 10 seconds needs more clips, and more clips mean more chances to pay for a retake.

Bottom line

A $0.05 a second headline is real, but it is a vendor list price for one model. On Sume the 720p range in the models priced here runs from $0.125 to $0.578 a second. Pick by what the clip needs, then multiply by its duration.