What one second of AI video actually costs

Model APIs and platform workbenches, converted to cost per second and put in one table. Prices come from vendor pricing pages and an October 2026 cross-platform test. Anything that can't be computed is marked —; the only estimates are the one-time GPU prices on the open-source page, labelled as such.

¥0.09/sec

Cheapest anywhere · Vidu Q4 launch price

Baseline ¥0.63 / sec

Comparison point · Seedance 1080P

online entries · 13 model API / 7 platform · plus 12 open source

Online models 20 Model API · 13 Platforms · 7 Open source · 12 Overseas · 7

Online models · 20Model APIs and platform workbenches, cheapest first; the 20 online quotes contain no estimates. The 12 open-source models have no per-second price, so their one-time GPU cost (an estimate that may rise) is shown on a separate page.

Range ¥0.09 – ¥5.05Spread 56×

Model layer · 11Official API keys billed per second. You assemble the workflow yourself.

Range ¥0.09 – ¥5.05Quotes 11

Platform layer · 6Credits converted to cost per second. Switching model inside one platform can change the price 4×.

Range ¥0.476 – ¥2.28Quotes 7

Open source · 12Free weights are not free hardware: you buy the GPU up front. One-time estimates, subject to price rises. Six of the twelve are Apache 2.0; the floor is a single 5090, the ceiling 8× H100.

Apache 2.0 6 / 12Hardware from ≈ ¥24,000

Overseas · 7Official per-second API prices from Google / Runway / Luma / xAI / Pika, converted at the 2026-11-10 PBoC mid rate (1 USD = ¥6.7167) so they sort in the same unit as the domestic rows; the USD list price stays in each subtitle. Subscription-only tools (Midjourney, Firefly) and the retired Sora 2 API are not included.

Range ¥0.134 – ¥0.806Spread 6×

Model layer Direct

Buy an API key or subscription straight from the vendor. The most transparent option, billed per second — but you assemble the script, storyboard and voiceover yourself. Currently as low as ¥0.09/sec.

Platform layer Packaged workflow

Models wrapped into a workbench: storyboards, character consistency, batch runs. Billed in credits, and switching models inside one platform can change the price 4×. You are paying for the workflow, not the pixels.

Open source Highest bar

Weights are free, hardware is not: you buy the GPU up front (one-time estimate, may rise) and still pay for power and time. Six of the twelve are Apache 2.0 (HappyHorse-1.0, SANA-WM, Bernini, MAGI-2, LingBot-Video, LongLive 2.0); LTX-2 and LTX-2.5 use community licences and SANA-Video uses the NVIDIA Open Model License. Hardware runs from one RTX 5090 to 8× H100. See all 12 →

DurationBudget ≤AnyVendor

Sorted by cost per second, cheapest first · click a column header to reverse Prices updated 2026-11-10 · 0 days ago

Price comparison table
#	Platform / model	Layer	Price▲	vs Seedance 1080P
1	APIBest valueVidu Q4 Preview ↗Shengshu Technology · launch pricing, released 2026-10-08	Model	¥0.09Cheapest anywhere	▼ ×0.14
2	APIJimeng · Seedance 720P ↗ByteDance / Volcano Engine · official per-second price	Model	¥0.28	▼ ×0.44
3	PLTPixmax ↗Ruiyun · Seedance 2.0 720P, pay as you go	Platform	¥0.476Cheapest on platform layer	▼ ×0.76
4	PLTSpecial Cat AI ⌕Seedance 2.0 Mini 720P, billed in cat-food credits	Platform	¥0.50	▼ ×0.79
5	APISeedance 2.0 Mini 720PReseller list price on a third-party platform	Model · reseller	¥0.50	▼ ×0.79
6	APIMiniMax-H3 768P ↗Hailuo / MiniMax	Model	¥0.50	▼ ×0.79
7	APIJimeng · Seedance 1080P ↗ByteDance / Volcano Engine	Model	¥0.63	= base
8	APIMiniMax-H3 2K ↗Hailuo / MiniMax	Model	¥0.80	▲ ×1.27
9	PLTupdream Pro ↗Bilibili · ¥159 / 19,800 credits, Seedance 2.0 720P	Platform	¥0.883Drops to ¥0.40 with MiniMax H3	▲ ×1.4
10	APIHappyHorse 1.1 720P ↗Alibaba ATH · KuaiLeMa	Model	¥0.90	▲ ×1.43
11	PLTSeko · first purchase ↗SenseTime · ¥39 / 500 credits, Seedance 2.0 720P	Platform	¥0.95Intro price — regular price below	▲ ×1.51
12	PLTSkylark ⌕Seedance 2.5 full 720P, measured over 240 sec	Platform	<¥1.00Cheapest full 2.5	▲ ×1.57
13	APIJimeng · Seedance Pro ↗ByteDance / Volcano Engine	Model	¥1.00	▲ ×1.59
14	APIHappyHorse 1.1 1080P ↗Alibaba ATH · KuaiLeMa	Model	¥1.20	▲ ×1.9
15	PLTSeko · regular ↗SenseTime · ¥60 / 500 credits, same spec	Platform	¥1.4650% more without the intro deal	▲ ×2.32
16	PLTTapNow ⌕Seedance 2.5 full 720P, among the priciest measured	Platform	¥2.28	▲ ×3.62
17	APISeedance 2.5 official retail ↗¥39 / 15 sec · bought per clip, not wholesale	Model · retail	¥2.60Costs more than most platforms	▲ ×4.13
18	APISeedance 2.0 4KTop resolution tier	Model	¥5.05Priciest in the table	▲ ×8.02
19	APIWan3.0-Video 720P ↗Alibaba Cloud Model Studio · official list ¥0.60/sec, 30% off through Nov	Model	¥0.60	▼ ×0.95
20	APIKling 4.0 ↗Kuaishou · 2026-10-21 API launch, third-party benchmark ¥0.90/sec	Model	¥0.90	▲ ×1.43

Nothing matched — try another keyword.

Methodology: how these numbers are computed

CNY/sec = official list price ÷ billed seconds; credit platforms = plan price ÷ seconds that tier can produce.
Overseas layer = USD list price × PBoC (CFETS) mid rate (currently 2026-11-10, 1 USD = ¥6.7167).
Open source has no per-second price — the cost is one-time hardware (GPU estimates, may rise) plus power and time, hence a separate page.
Subscription-only tools (Midjourney, Firefly) and retired APIs are not collected; anything computable is computed. The only estimates are the open-source GPU prices, labelled as such and subject to change — check live listings before buying.

Compare the spec before the price. A per-second figure only compares when it is the same model at the same resolution. Pixmax, Special Cat, updream and Seko are normalised to Seedance 2.0 / 720P; Skylark, TapNow and the Seedance 2.5 retail row are Seedance 2.5 full 720P — the two groups are not interchangeable. Most credit-based platforms never publish per-clip consumption, so those rows use subscription fee ÷ seconds that tier can produce.

Three traps worth knowing. ① Official retail is not the cheapest — Seedance 2.5 at ¥39 / 15 sec (¥2.6/sec) costs more than most platform annual plans. Platforms subsidise with annual prepay to lock you in. ② Open source ≠ commercially usable — six of the twelve are Apache 2.0, but LTX-2 / LTX-2.5 and MiniMax H3 use community licences (H3 needs a separate application in the US/EU/UK/KR), and SANA-Video uses the NVIDIA Open Model License, so read the terms before you ship. ③ Open source ≠ free — the weights cost nothing, but the GPU is a one-time purchase (estimated here, and prices can rise), plus power and your time.

⚖

Data on this page is indicative only — listing a vendor is not a recommendation or guarantee.Figures are compiled and converted from public pages and can change at any time; check the vendor's current pricing before you buy. Read the disclaimer