Cheapest 720p AI video per second: Veo, xAI, Sume

The cheapest 720p per-second rate in this comparison is Sume's Grok Imagine Video 1.5 at $0.0125, against $0.05 for Veo 3.1 Lite and $0.05 for xAI's listed grok-imagine-video. The most expensive is Veo 3.1 Standard at $0.40. A 10-second clip ranges from $0.125 to $4.00.

Ranked

Vendor rows are list prices; Sume rows are provider list times 1.25.

Careful with the first rows

Sume's Grok Imagine Video 1.5 and the xAI model listed as grok-imagine-video may not be the same model or tier. The xAI page does not say 720p for its single rate, so match output specs before you treat them as equals.

The price difference between Grok and Wan on Sume is ten times; that gap is larger than any gap between Sume and a vendor's list price for the same model.

• Cheapest does not mean best at motion, text or audio.
• Sume Grok accepts 4 to 15 seconds per request and is image-to-video only, so it needs a start image; the other rows also take text-only prompts.
• Veo Lite lists 720p and 1080p, with no 4K tier.

What a 100-clip test costs

If you want to compare quality before choosing, 100 ten-second clips at 720p cost $12.50 on Sume Grok, $50.00 on Veo Lite, $100.00 on Veo Fast, $125.00 on Wan or Omni Flash and $400.00 on Veo Standard. Even a small side-by-side of 10 clips per option stays at or under $40 on every row.

Run the same ten prompts on every candidate, rank them blind, then price the winner at your full volume. The price table above is where that last step starts.

Limits

Rates and model names change; Veo and xAI prices are public list prices and exclude any volume terms. The tables show the model amount. A workspace Agent Fee line can be added on top, so treat the job's own usage cost as the truth and check the live catalog before you budget; limits and rows change.