AI video price per minute table, next to Sume's per-model SKUs

A Hedra comparison citing Artificial Analysis lists AI video at $6.00 a minute for Gemini Omni Flash at 1080p, up to $20.16 for Kling V3 Pro. Sume prints its own price for each model in pricing_skus, from GET /v1/videos/models, so you can compare on the same unit yourself.

The table

All rows come from Hedra's table (listed at 1080p default unless a row says otherwise), which cites an Artificial Analysis snapshot dated July 17, 2026. Prices change, so treat this as a dated reference.

Turning a minute into a clip

A per-minute price is easy to convert. Divide by 60 for a per-second figure, so $6.00 a minute is $0.10 a second, and multiply by the seconds you need. A 10 second clip at that rate is $1.00. Those two numbers are arithmetic on the table, not a quote.

Remember that you rarely keep the first generation. A budget should include retries, and a per-minute price hides how many attempts a usable second takes.

Where Sume prints its price

Every model in Sume's video catalog carries a pricing_skus object in GET /v1/videos/models. The docs example shows one keyed by a unit such as per-1000-video-tokens, so units differ by model. Billing is the provider list price times 1.25. The amount reserved on submit comes out of your workspace balance, and usage.cost on the finished job is the billed amount.

Sume lists Gemini Omni Flash 1.1 as gemini-omni-flash-1.1, Seedance as seedance-2 and seedance-2.5, and others by their own ids. This post does not repeat a Sume price, because the catalog is the source and it can change.

curl "https://api.sume.com/v1/videos/models" \
  -H "Authorization: Bearer $SUME_API_KEY"

Compare on the same terms

Match resolution, audio and duration before you compare. A 4K row against a 1080p row is not a like-for-like comparison, and a model that always includes native audio is priced differently from one with an audio switch.