The third billing model nobody mentions

Two billing shapes are familiar: a figure in currency, or a credit rate with a credit price. A third exists, where output is priced in tokens computed from resolution and duration, and a tier change can multiply the same clip by three. As of 2026-09-12.

Inclusion rule. Tiers whose vendor publishes a token rate per second. A tier priced only in currency belongs on the ledger rather than here, and the relationships shown are proportional rather than absolute. Order. Alphabetical by output tier.

1Why tokens change how a tier is chosen

With a currency rate, moving up a resolution multiplies the bill by a published ratio. With tokens, the multiplier is computed from the output and can differ from what the tier table suggests, because duration and resolution both feed the same number.

The practical consequence is that the cheapest tier is not always cheapest per finished minute, and working it out needs the token formula rather than a price list.

It also changes what a test generation proves. A five-second clip at the tier a season will actually ship at is a usable estimate; the same clip at a cheaper tier is not, because the token count moves with the output rather than tracking a fixed ratio between tiers.

2Multipliers inside one platform reach three times

The spread between the lowest and highest output tier on a single platform can be a factor of three. That is larger than the gap between many channels, which means the tier decision matters more than the channel decision for a production that has already settled on a resolution.

The relationship with duration is more regular: the tiers above reach the same multiple between a five-second and a fifteen-second clip. Length behaves the way intuition expects. Resolution is where the arithmetic surprises people.

3Six things to check before trusting any per-second figure

Whether credits expire. Whether failed generations are refunded. Whether there is a minimum billable duration. How partial seconds are rounded. Whether any tier carries a surcharge. Whether the channel bills per returned second or per requested second.

Each of those moves the effective rate, and none of them appear in a headline figure. Two channels with identical published rates can differ by a fifth once all six are settled.

Rounding is the quietest of the six. A channel that rounds up to whole seconds charges a season of six-and-a-half-second shots as if every shot were seven, and nothing on the pricing page says so.

4What the ledger records and what it refuses

Figures that can be read in public, at a stated resolution. A token rate without the token price, or a credit rate without the credit price, is not a rate and is recorded as unpublished rather than converted on a guess. The rule is on the method page.

5The production behind the price

Price aside, what goes into the seconds behind “Token billing” is described in SceneMixer's guides:

Video is 16:9 landscape by default, with 9:16 vertical available; you set it once per project. Segments are typically 4 to 15 seconds, and some model tiers go to 30.

SceneMixer, AI Short Drama Generator — “What length and aspect ratio does it output?” (read 2026-10-13)

A segment is one generation, and its length is the sum of its shots’ seconds.

SceneMixer, How to Change a Segment’s Duration, and Which Tiers Accept 30 Seconds (read 2026-10-13)

6Where the sourced entries live

This page is about how the pricing works. What any particular channel charges is a separate matter, kept in the ledger with the page each figure came from and the date it was read.

• BytePlus ModelArk documentation — publishes a token rate per second of output
• MiniMax pay-as-you-go pricing — publishes pay-as-you-go rates per generation

Written to explain, not to record. No sentence here is attributed to a seller, and none of it should be quoted as a fact about one. The sourced material is on the ledger. Related: Resale and source, Delivery resolution.