AI Video API Cost in 2026: List Price vs Usable-Clip Cost

Nobody overspends on video generation because they picked the wrong rate card. They overspend because they budgeted for the clips they wanted and paid for the clips they threw away. The per-second rate is the term everyone compares and the smallest term in the bill. The term that actually moved in 2026 is how many generations it takes to get one shot you can cut into a timeline, and that number roughly tripled without a single price change.

The formula:  effective cost = list rate x seconds x attempts per usable clip
What moved:   attempts per keeper went 2-3 -> 7-8 during 2026 congestion (36kr, 2026-04-01)
What did not: the rate card
Measured:     6s 480p Seedance 2.5 billed $0.66 in 103s; 30s billed $3.30 in 200s
Delivered:    that same 6s clip is $1.32 at 2 attempts and $5.28 at 8
Expiring:     the 1080p launch discount ends 2026-09-17 on every gateway that mirrored it
Snapshot:     2026-08-25

Everything below is either a rate card we read on 2026-08-25, a generation we ran and were billed for, or reporting we cite by name. Where the number people repeat does not survive a check, we say so rather than repeating it.

Why Is the Rate Card the Wrong Number?

Because it prices generations, and you ship keepers.

Every video API bills per output second of whatever comes back, including the takes you delete. So the unit that matters is not the per-second rate but the per-second rate multiplied by your reject ratio:

effective_cost_per_delivered_second = list_rate x (generations / accepted)

That second factor is the one nobody puts in a pricing comparison, because no vendor can quote it. It depends on the model, your prompt discipline, how specific your brief is, and, uncomfortably, on how busy the provider is that week.

Here is what that does to a real bill. We generated a 6-second 480p clip on Seedance 2.5 on 2026-08-13 and were billed $0.66:

The spread between the top and bottom of that table is 8x. The spread between the cheapest and most expensive rate card on the market for the same model is under 3x. If you are optimising the rate card while ignoring the attempt count, you are tuning the smaller dial.

What Actually Changed in 2026?

The attempt count, and it changed for reasons outside your pipeline.

This is the part with real reporting behind it. In April, 36kr documented what Seedance users were seeing: where creators had previously generated two or three clips and picked one usable take, they were now often generating seven or eight to find one. Same prompts, same model name, roughly triple the burn.

The surrounding conditions were not subtle. That same reporting describes queue numbers reaching into six figures. 21jingji reported queue positions climbing from 1,900 to 50,000 and waits stretching to 68 hours, alongside capacity management that Chinese coverage named 降智, roughly "dumbing down": throttling output quality on some task types to clear the backlog. Consumer-side pricing moved at the same time, with early-adopter discounts withdrawn and the cost of a 15-second generation for paying members rising about 167%.

You do not have to accept any editorial framing to use this. The operational lesson stands on its own: your provider's capacity state is an input to your unit economics, and it is not on the rate card. A quarter where attempts per keeper drifts from 3 to 6 doubles your content cost while your finance dashboard shows a flat price per second.

That is also why a cost model with a single hardcoded multiplier is worthless. Instrument the ratio and recompute it weekly.

What Do the Rate Cards Actually Say?

Read on 2026-08-25, for the same model, so the comparison is real.

Four things in that table are worth more than the cheapest cell.

The first is that the same model is not the same product on every route. ModelArk and Volcano Engine Ark are separate consoles with separate model IDs, dreamina-seedance-2-5-260628 internationally and doubao-seedance-2.5 in China, and you cannot mix credentials or IDs between them.

The second is that credit pricing is rate pricing wearing a costume. Runway's API credit is one cent, so 30 credits per second is thirty cents per second, before the surcharge it adds for each second of input video you supply. Higgsfield's credits only convert through a subscription, which means its effective rate depends on how much of your monthly allowance you actually burn. A plan you use at 40% is a rate card 2.5x worse than the one advertised.

The third is that token billing and per-second billing answer different questions. ByteDance bills its own API per token, which is honest to how the model works and useless for forecasting a 25-shot episode until you have run the token formula against real prompts. Per-second gateway pricing is a markup on that, bought for predictability.

The fourth is the one with a date attached, and it is below.

What Did We Actually Get Billed?

Three runs on 2026-08-13, and the numbers are less flattering than the rate card implies.

The last row is the good news and it is worth building for. When Seedance 2.5 was still capped at 720p, a 1080p request returned resolution 1080p not supported; allowed: [480p 720p] rather than quietly downscaling and billing you. A synchronous 400 costs nothing. Any parameter you can get rejected at submit time is a parameter that cannot become a wasted generation, so validate aggressively and prefer providers that fail loudly.

The 2.0 versus 2.5 pair is the sharper lesson. Seedance 2.0 bills 36% less per second at 480p and took 2.8x longer to return. If 2.0 also needs one extra attempt on your shot list, its price advantage is gone and you have paid for the privilege in latency. We wrote up where each model still wins in Seedance 2.5 vs 2.0.

How Do You Actually Cut the Bill?

Attack attempts. The rate card has maybe 30% of movement in it; the attempt count has 400%.

Four levers, in the order we would apply them to an existing pipeline.

Draft cheap, deliver expensive. At $0.11 against $0.48, 480p costs less than a quarter of 1080p on the same model. Iterate the composition, the beats and the camera at 480p, then re-render only the approved shot at delivery resolution. A 25-shot episode that iterates four times at 480p and renders once at 1080p costs meaningfully less than one that iterates twice at 1080p, and it is the same finished episode.

Remove the model's freedom to reinterpret. Every degree of freedom you leave open is a coin flip you pay for. Reference assets pin identity and setting; Seedance 2.5 accepts up to 50 of them as 30 images, 10 videos and 10 audio clips. First and last frames pin where a shot starts and ends, and in our first and last frame test the 2.5 last-frame error was 5.4 against 35.9 on 2.0 Mini. Integer-second timestamps pin when beats land. None of these lower the rate. All of them lower how often you re-roll.

Get your task type right the first time. On Seedance 2.5 the task is inferred from words in your prompt, so an edit request without an edit trigger word is classified as a reference-to-video job and returns a brand new clip. It is a perfectly good clip. It is also a full billed generation that answered a question you did not ask. Setting the task type explicitly moves that failure to submit time, where it is free. The trigger words and the failure modes are in how to use Seedance 2.5.

Measure the ratio, not the average. Log billed amount and generation count per accepted shot, not per request. A per-request average tells you the rate card, which you already knew. Cost per accepted shot tells you when your provider's capacity state has quietly moved, usually two weeks before anyone notices in the finance review. If you are polling asynchronously, the same log gives you queue behaviour for free; we broke that pattern down in video generation API polling.

What Expires on 2026-09-17?

The 1080p price, everywhere.

Seedance 2.5 shipped without 1080p and gained it on 2026-08-17 Beijing time. ByteDance attached a launch discount running to 2026-09-17, and the gateways mirrored it at various depths: 20% off at ofox and Atlas Cloud, 28% off at EvoLink, 72% of list on Volcano Engine Ark. Different numbers, one shared expiry date.

If you are building a cost model this month, carry both the promotional and the post-promotional 1080p rate, because a spreadsheet anchored on $0.48 per second silently becomes wrong on September 17th and nothing in your pipeline will tell you. On a 25-shot, 60-second episode delivered at 1080p, the difference between $0.48 and $0.60 per second is $7.20 per episode before attempts, and $28.80 to $57.60 after a realistic multiplier. At 1,000 episodes that is a budget line, not a rounding error.

Three Numbers We Could Not Verify

Cost claims travel further than their sources, and this topic has three in wide circulation that we could not stand behind, so we are not repeating them as facts.

A widely quoted V2EX post said to be titled along the lines of "it's not the model, it's your budget" does not appear to exist under that title. The closest real thread, a June discussion of pricing, failure and waiting, argues that unpredictable pricing hurts more than model quality does, which is a fair point, but it contains no cost figures at all. The commonly repeated per-clip and per-ad figures attached to it have no traceable origin.

A per-episode figure of roughly $7.6 for a 60-second, 25-shot production is often attributed to Tencent Cloud. Tencent Cloud's own pipeline write-ups publish cost reduction percentages and per-task unit costs, such as storyboard splitting at CNY 2.8 against CNY 9, and production timings of about three hours per 2-3 minute episode. They do not publish a total per-episode cost. The $7.6 may be someone's arithmetic; it is not a vendor figure.

And the queue number frequently cited as 80,000 is not what the reporting says. 36kr describes queue numbers reaching around 100,000; 21jingji describes a rise from 1,900 to 50,000. Both are worse than 80,000 in different ways, and neither is 80,000.

If you are building a business case on any of these, build it on your own logs instead. The one number that matters is one only your pipeline can produce.

The Short Version

The rate card is a starting point that vendors compete on and a term you have almost no leverage over. Attempts per usable clip is the term that moved 3x in 2026, that varies with your provider's capacity state, that responds to reference assets and explicit task types and resolution laddering, and that nobody will quote you.

Price it, log it, and recompute it weekly. Then compare rate cards, with the multiplier already in the sheet.

Frequently Asked Questions

How much does an AI video generation API actually cost per second?

List rates for Seedance 2.5 on a gateway run $0.11 per output second at 480p, $0.24 at 720p, and $0.48 at 1080p during the launch discount window. Runway's API bills the same model at 20 credits per second at 480p and 30 at 720p, which is $0.20 and $0.30 at the API's one-cent credit. Those are the numbers on the rate card. What you actually pay per clip you can ship is that rate multiplied by how many generations you burn to get one keeper, which in 2026 has been anywhere from 2 to 8.

What is the effective cost of an AI video clip?

Effective cost is the list rate times the number of generations per usable output. A 6-second 480p Seedance 2.5 clip billed us $0.66. At the 2-3 attempts per keeper creators reported before the 2026 congestion, that clip costs $1.32 to $1.98 delivered. At the 7-8 attempts 36kr reported during the worst of it, the same clip costs $4.62 to $5.28. The rate card did not move. The bill tripled.

Why do I need so many generations to get one usable clip?

Three causes, and only one of them is your prompt. Capacity throttling degrades output quality during congestion, which Chinese coverage in 2026 labelled 降智. Underspecified prompts leave the model free to choose, and it chooses differently every seed. And ambiguous task typing sends an edit request down a reference-to-video path, so you get a technically fine clip that answers the wrong question. The second and third are fixable in your pipeline. The first is not.

Is a cheaper per-second model actually cheaper?

Only if its attempt count holds. Seedance 2.0 at 480p bills $0.07 per second against $0.11 for 2.5, so 2.0 looks 36% cheaper. If 2.0 needs four attempts for a shot that 2.5 lands in two, 2.0 costs more per delivered second and takes longer. We measured 289 seconds of wall clock on 2.0 against 103 on 2.5 for the same 6-second prompt, so the slower model also spends more of your pipeline's time.

How do I lower video API costs without changing models?

Cut attempts, not rate. Draft on 480p and only re-render the approved shot at delivery resolution, since 480p is a fifth of the 1080p rate. Pin the parts you are not iterating on with reference assets and first and last frames. Write integer-second beats so the model is not inventing structure. And log the billed amount per accepted shot rather than per request, because a per-request average hides exactly the number you need.

Does the 1080p discount on Seedance 2.5 last?

No. Seedance 2.5 gained 1080p on 2026-08-17 Beijing time, and ByteDance attached a launch discount that runs to 2026-09-17. Gateways mirrored it at varying depth, from 20% off to 28% off. Every one of those prices reverts on the same date, so a cost model built on the promotional 1080p rate breaks in September. Model both numbers now.

Originally published on ofox.ai/blog.