Vidu Q4 540P at $0.045/s vs Sume's cheapest rows: $0.0375 to $0.0625

The short answer

Vidu Q4 Preview starts at $0.045 per second at 540P, according to its launch release. Sume does not list Vidu, so you cannot buy that rate on Sume. The nearest low-end Sume rows are Gemini Omni Flash 1.1 at 360p ($0.0375 per second) and the 480p tiers of MiniMax H3, MiniMax H3 Max and Wan 3.0 (all $0.0625 per second). Vidu figures come from the launch press release; Sume figures come from the public catalog and docs. Sume does not list Vidu, so every Sume row below is a different model.

Resolution is not the same across these rows, so the table is a price floor comparison, not a quality comparison.

Cheapest Sume rows against the Vidu starting price

Sume bills the provider list price times 1.25. The figures below are Sume prices after that margin. The Vidu figure is the vendor's own rate. The last column divides each Sume rate by $0.045.

What a 10-second clip costs

At 10 seconds the arithmetic is straightforward: Vidu 540P is 10 x $0.045 = $0.45. Omni 360p is 10 x $0.0375 = $0.375. Wan 3.0 at 480p is 10 x $0.0625 = $0.625. Seedance 2.0 at 480p is 10 x $0.17577 = $1.7577.

Only one Sume row is below the Vidu starting price, and it is the lowest resolution Omni offers. If you need 540P-class output, the nearest step up on Sume is the 480p tier of Wan 3.0 or MiniMax, at 39% more per second than Vidu's starting rate.

The discount changes the floor

The release also says image-to-video and reference-to-video are 30% off through November 30. At 540P that is $0.045 x 0.70 = $0.0315 per second, below every Sume row in the table. The discount is time-limited, so a budget built on it needs a fallback. The Vidu Q4 test-batch planner covers that case.

How to decide

Price per second only matters if the clip is usable. Sume's video generation docs show each model's supported resolutions, durations and reference types through GET /v1/videos/models, so you can check a row before you spend on it. Run one short probe per candidate row, keep the prompt identical, and compare the clips side by side before you scale. For a first-pass draft where framing matters more than finish, a cheap row is enough. For a delivered clip, compare the output and not the rate. To see whether Vidu has since been added, read the catalog check.