On this page
For the lowest 720p price with audio, call Veo 3.1 Lite on Vertex AI ($0.05/s). For one credit balance across makers, use the Runway API ($0.01 per credit). For Seedance, go to BytePlus rather than a reseller, which charges about twice as much. Plan for shutdowns: the Sora API and the Veo 3.1 Gemini API preview IDs have already been switched off.
Last checked . Prices are re-checked every week (planned: Wednesdays). Each row shows its own check date; a row older than 14 days is out of date.
How the APIs bill
| API | Billing unit | Notes |
|---|---|---|
| Google Vertex AI (Veo 3.1) | per second of output, by tier, resolution and audio | us-central1 only per the model reference; outputs carry SynthID |
| Gemini API (Gemini Omni 1.1 Flash) | output tokens: $17.50 per 1M, 5,792 tokens per second at 720p | GA since 2026-08-27; 3 to 10 s clips, extension up to 40 s total |
| Kling API | units per second, sold as prepaid packs | flagship packs from $500, general from $700, 180-day validity, no rollover |
| BytePlus ModelArk (Seedance) | tokens: seconds x pixels x fps / 1024 | online and cheaper offline (batch) inference on some models |
| Runway API | credits at $0.01, per second | resells Veo, Kling, Seedance, Wan, MiniMax, Grok, Gemini Omni |
| Luma API (Ray3.2) | per clip: 5 s or 10 s | HDR 2x, HDR+EXR 3x, 10 s clips 3x the 5 s price |
| MiniMax API (H3 family) | per output second | input video billed per second too |
| Alibaba Model Studio (Wan, HappyHorse) | per second of successful output | failed requests free; Wan 3.0 bills input plus output seconds |
| xAI API (Grok Imagine) | per second by resolution | input images $0.002 to $0.01 each, input video $0.01 per second |
| LTX API | per second by resolution | Studio subscriptions do not include API access |
| fal.ai, Replicate | per second, per clip or per token | pay as you go, see reseller comparison |
Price per second at 720p
| Model | Route | Basis | Per second | Per clip | Flags | Source | Checked |
|---|---|---|---|---|---|---|---|
| Seedance 1.0 Pro Fast Seedance | API BytePlus ModelArk | 720p 16:9 (derived: 102,960 tokens x $1/M / 5 s) | $0.0206/s | - | derived | source | Nov 5, 2026 |
| Veo 3.1 Lite Veo | API Vertex AI | 720p no audio silent | $0.03/s | - | preview | source | Nov 5, 2026 |
| Grok Imagine Video 1.5 Lite Grok Imagine | API xAI API | 720p | $0.03/s | - | - | source | Nov 5, 2026 |
| LTX-2.3 Fast LTX | API LTX API | 720p | $0.03/s | - | - | source | Nov 5, 2026 |
| LTX-2.3 Pro LTX | API LTX API | 720p | $0.04/s | - | - | source | Nov 5, 2026 |
| Veo 3.1 Lite Veo | API Vertex AI | 720p with audio audio | $0.05/s | - | preview | source | Nov 5, 2026 |
| Seedance 1.0 Pro Seedance | API BytePlus ModelArk | 720p 16:9 (5 s = $0.26) | $0.052/s | - | - | source | Nov 5, 2026 |
| Ray3.2 Luma Ray | API Luma API | 720p SDR 5 s ($0.30) | $0.06/s | $0.30 / 5 s | derived | source | Nov 5, 2026 |
| Kling 4.0 Flash Kling | API Kling API | 720p, with input video | $0.0615/s | - | derived | source | Nov 5, 2026 |
| Grok Imagine Video (1.0) Grok Imagine | API xAI API | 720p | $0.07/s | - | - | source | Nov 5, 2026 |
| Veo 3.1 Fast Veo | API Vertex AI | 720p no audio silent | $0.08/s | - | - | source | Nov 5, 2026 |
| Seedance 2.0 Mini Seedance | API BytePlus ModelArk | 720p | $0.08/s | - | - | source | Nov 5, 2026 |
| Kling 3.0 Kling | API Kling API | 720p no audio silent | $0.084/s | - | derived | source | Nov 5, 2026 |
| Kling 3.0 Omni Kling | API Kling API | 720p no audio silent | $0.084/s | - | derived | source | Nov 5, 2026 |
Rows marked derived are computed from published tokens, units or credits. The numbers are list prices; enterprise discounts and committed-use deals are not tracked.
A month of usable clips
A budget view: 1,000 usable 8 s clips need 2,000 generations at a 50% keeper rate (an assumption you should replace with your own), or 16,000 seconds of output.
| Route, 720p | Per second | Monthly bill |
|---|---|---|
| Veo 3.1 Lite, Vertex, with audio | $0.05/s | $800 |
| Veo 3.1 Fast, Vertex, with audio | $0.10/s | $1,600 |
| Gemini Omni 1.1 Flash, Gemini API | $0.101/s | $1,616 |
| Runway Gen-4.5 | $0.12/s | $1,920 |
| Kling 3.0, with audio | $0.126/s | $2,016 |
| Seedance 2.0, BytePlus | $0.15/s | $2,400 |
| Seedance 2.0, fal.ai | $0.303/s | $4,854 |
| Veo 3.1 Standard, Vertex, with audio | $0.40/s | $6,400 |
Different models, same row of arithmetic: the bill moves with the keeper rate. Halve it and every figure doubles. Use the calculator with your own numbers.
Picks by need
Lowest cost with audio: Veo 3.1 Lite on Vertex
$0.05/s at 720p, $0.08/s at 1080p. It is a preview model with no 4K price and no reference-image support, and Vertex lists us-central1 only. Clips up to 8 s. See Veo 3.1 pricing.
Gemini API: Gemini Omni 1.1 Flash
The only Google video model sold on the Gemini API after the Veo 3.1 preview IDs were shut down on 2026-10-22. Google states about $0.101/s at 720p, token-billed, with 3 to 10 s clips.
One bill for many models: Runway API
Credits at $0.01, per second. It lists Gen-4.5 at $0.12/s, Gen-4 Turbo at $0.05/s and Aleph 2.0 (video editing) at $0.28/s with a 56-credit minimum, plus resold models at or above first-party prices. See Kling vs Runway price.
Kling: the direct API for predictable units
Kling 3.0 is $0.126/s with audio, up to 15 s per the Replicate page. The prepaid structure suits steady usage and punishes occasional use. See Kling API cost per second.
Seedance: BytePlus first-party
Token billing, so price tracks pixels. 720p is $0.15/s for 2.0 and $0.231/s for 2.5. Offline inference is listed at $4.00 per million tokens for 2.0 at 1080p against $7.00 online, for jobs that can wait. Seedance 2.0 on fal.ai is about double. See Seedance pricing.
Predictable per-clip pricing: Luma API
Ray3.2 is $0.30 for a 5 s 720p clip and $0.90 for 10 s, so you know the cost of a call before it runs. Per second it is $0.06/s on 5 s clips.
Shutdown and deprecation risk
Developer pages rarely mention that endpoints disappear. In this tracker, four have been retired or dropped:
- Sora 2, Sora 2 Pro and the Videos API: OpenAI announced the shutdown on 2026-03-24, effective 2026-09-24, with no replacement. See Sora API alternatives.
- Veo 3.1, 3.1 Fast and 3.1 Lite preview IDs on the Gemini API: shut down 2026-10-22; Gemini Omni 1.1 Flash is the replacement. Veo 2 and Veo 3.0 IDs on the Gemini API went on 2026-06-30, announced 2026-06-15.
- LTX-2 model IDs: unsupported from 2026-08-15, with requests routed to LTX-2.3 at LTX-2 rates.
- Kling 2.5: no longer listed on Kling's own API page.
- Veo 3.1 Lite on Vertex is labelled preview today, so its terms may still change.
Practical steps that follow from this: pin the model ID, keep a second route (a reseller or a second maker) wired up and priced, read each provider's deprecations page on a schedule, and avoid long prepaid packs for a model that may be retired inside the validity window. The price history lists dated events.
What is not tracked
Rate limits, queue times, uptime, SDK quality and enterprise discounts are not in the dataset. Max clip length is filled only where a source states it (see length limits). Check the commercial-use terms of each API before shipping: commercial use rules.
Which API?
- Pick Veo 3.1 Lite or Fast on Vertex for the lowest 720p with audio price and a first-party, non-reseller route.
- Pick the Runway API when you want several makers behind one balance and can accept first-party-or-higher prices.
- Pick BytePlus for Seedance, Kling for steady Kling volume, Luma for fixed per-clip costs.
- Pick fal.ai or Replicate when you need pay-as-you-go without a prepaid pack, and choose models at parity price.
- Whatever you pick, wire a fallback route.
Questions people ask
What is the best AI video generator API?
It depends on cost and model. On price alone, Veo 3.1 Lite on Vertex is lowest at 720p with audio; for several models on one bill, the Runway API.
How much does the Veo 3 API cost?
Veo 3.1 on Vertex: $0.40/s Standard, $0.10/s Fast, $0.05/s Lite at 720p with audio. The Gemini API Veo IDs are shut down.
How much does the Runway API cost?
$0.01 per credit: Gen-4 Turbo 5 credits per second, Gen-4.5 12, Aleph 2.0 28.
What does the Seedance API cost?
Token billing on BytePlus: $0.15/s for 2.0 at 720p and $0.231/s for 2.5.
Is there a free AI video generator API?
No free API tier is tracked here. For free app tiers see free AI video generators.
Sources
Prices and limits were read from the sources below on the dates shown. They change often: confirm on the provider's page before you pay.
- Google Cloud Veo pricing - 2026-11-05
- Gemini API pricing - 2026-11-05
- Gemini API deprecations - 2026-11-05
- OpenAI API deprecations - 2026-11-05
- Kling API pricing - 2026-11-05
- BytePlus ModelArk pricing - 2026-11-05
- Runway API pricing and costs - 2026-11-05
- Luma API pricing - 2026-11-05
- LTX API pricing - 2026-11-05