Short answer: The most direct way to get a Seedance 2.5 API is the Seedance 2.5 API at /docs/api: Bearer-key REST endpoints, a free price quote before every generation, and the same credits and prices as the website, with Seedance 2.5 Lite at 10 credits per output second (about $0.035–$0.05 per second on annual plans). A second Seedance 2.5 API with its own docs works the same way and adds multimodal references. Among third-party providers, BytePlus ModelArk is the official token-billed platform, Replicate and OpenRouter list the lowest published 480p rate ($0.1028 per second), fal publishes clear token math, and Atlas Cloud and WaveSpeedAI quote simple "from" prices.
Integrating Seedance 2.0 models? Read our Seedance 2.0 API models and pricing guide.
Seedance API providers compared
| Provider | 480p (per second) | 720p (per second) | 1080p | How it bills | Main strength |
|---|---|---|---|---|---|
| Seedance 2.5 API (/docs/api) | Seedance 2.5 Lite: 10 credits per output second at 720p (100 credits ≈ $0.35–$0.50 on annual plans); Seedance 2.5: quoted per request via POST /quotes | Supported by Seedance 2.5; price quoted per request | Credits from your website balance; free quote before charging, refund on verified failure | Same balance as the web generator, max_cost_credits price cap, 4–30 second clips | |
| Seedance 2.5 API (multimodal docs) | Seedance 2.5 Lite: 10 credits per output second; Seedance 2.5: same pricing as the web generator, reported in billing.cost_credits | Supported by Seedance 2.5 | Credits from the shared account balance, with failure refunds | Text, image and multimodal (video/audio reference) inputs, up to 50 reference files | |
| BytePlus ModelArk | Not listed on official page (token-based) | Not listed on official page (token-based) | $11.7 per M tokens (no video input) | Per million tokens: $10.70 for 480p/720p without video input, $6.40 with video input | Official platform with a published 1080p token rate |
| Replicate | $0.1028 | $0.2312 | Not listed on official page | Per second of output; $0.4304 (480p) and $0.9676 (720p) when the input includes video | Simple per-second billing, lowest published 480p rate among resellers |
| OpenRouter | $0.1028 ("from" price) | $0.23112 | Not listed on official page | $10.70 per M tokens, passed through from the Seed provider | One OpenAI-compatible key for many models |
| fal | ~$0.2205 | ~$0.4730 | ~$1.164 (pricing note) | $0.0214 per 1,000 tokens at 480p/720p; ~$0.0234 at 1080p | Published token formula and worked cost examples |
| Atlas Cloud | From $0.134 per second (per-resolution rates not listed on official page) | Not listed on official page | Per second of generated video | One flat "from" rate across text, image and reference endpoints | |
| WaveSpeedAI | $0.90 per 5 seconds | $1.80 per 5 seconds | Not listed on official page | Per output, prorated per second from 4 to 30 seconds | Most endpoints, including turbo, video-extend and video-edit |
For rate limits, queue times and how these prices relate to the model's own limits, see our Seedance 2.5 API pricing and limits guide.
Alongside the provider documentation, review Seedance 2.5 and prepare a common test brief so your comparison reflects the workflow you intend to build.
Integrating in Python against our REST API first? Use the video generation API Python guide.
How does the Seedance 2.5 API at /docs/api work?
The Seedance 2.5 API documentation describes a REST API at the base URL /api/v1 with an OpenAPI 3.1 spec you can import into Postman, Insomnia or a client generator. It runs in three steps:
- Create a key. Make a key in Settings → API keys and send it as
Authorization: Bearer YOUR_API_KEY. Keys are server-side only; requests from browser code are rejected. - Quote, then submit.
POST /quotesreturnscost_creditswithout generating or charging. Then callPOST /videoswith a model ID such asseedance-2.5orseedance-2.5-lite, a prompt, and resolution, aspect ratio and duration. Addmax_cost_creditsto reject any request that would cost more than you expect. - Poll and download. A 202 response means the job is accepted. Poll
GET /generations/{id}no more than once every 10 seconds until the status issuccess, then download fromoutputs[].url. Each signed URL lasts up to 15 minutes, and media is kept for 14 days.
The API uses the same balance and pricing as the website. Generation requires an active paid subscription or remaining purchased credits, and credits are reserved before submission and refunded once if a generation fails. This version supports text-to-video, image-to-video, text-to-image and image-to-image; video and audio references, editing, webhooks and cancellation are not available. Each account can run one active API generation at a time, with up to 10 generation requests per minute and a 2,000-credit daily API budget.
Is there another Seedance 2.5 API with multimodal inputs?
Yes. The Seedance 2.5 API with multimodal references uses the same pattern: Bearer keys (sd_live_...), POST /videos, polling a returned status_url, and results in output.video_url. Its docs list text-to-video, image-to-video and multimodal-to-video modes. For Seedance 2.5 they list 480p, 720p and 1080p, whole-second durations from 4 to 30 seconds, and up to 30 images, 10 videos and 10 audios (50 files in total) as references. It shares the web account's credit balance and pricing, refunds failed generations, and asks for an Idempotency-Key on every create request. Media inputs must be public HTTPS URLs on approved storage hosts, as there is no upload endpoint yet.
Most users integrate on their own by following the /docs/api documentation; if you need more permissions or have special requirements, contact support.
Choose the /docs/api version for quotes before charging and a hard price cap; choose the multimodal version if you need video or audio references.
Which Seedance API provider is cheapest?
For Seedance 2.5 Lite at 720p, the Seedance 2.5 API has the lowest published rate here: 10 credits per output second, which is about $0.035 to $0.05 per second when credits come from an annual plan (100 credits ≈ $0.35–$0.50 on the pricing page). That is a Lite rate; full Seedance 2.5 prices depend on settings and are returned by a free quote.
Among third-party resellers at 480p and 720p without video input, Replicate lists the lowest per-second rates: $0.1028 at 480p and $0.2312 at 720p. OpenRouter shows a "from $0.1028/second" price and $0.23112 per second at 720p. fal is the most expensive per second, at roughly $0.2205 (480p) and $0.4730 (720p), and its own page says "Per-second figures are approximations for the common 16:9 case. The token formula is authoritative."
Watch the video-input surcharge. Replicate's 720p price rises from $0.2312 to $0.9676 per second when the input includes video, and fal counts input video duration in its token formula. To turn these rates into a cost per finished clip, use our Seedance 2.5 cost per video breakdown.
Which provider supports 1080p?
Both Seedance 2.5 APIs support 1080p for the Seedance 2.5 model, priced per request from the same rates as the web generator. Among third parties, BytePlus ModelArk's official pricing page lists a 1080p rate of $11.7 per million tokens without video input and $7.0 with video input. fal's pricing note mentions roughly $1.164 per second for 1080p, although its text-to-video parameter table lists only "480p" or "720p." Replicate, OpenRouter, Atlas Cloud and WaveSpeedAI do not list a 1080p price on their official model pages.
Which third-party provider fits which team?
- BytePlus ModelArk is the source platform. Its strengths are official token pricing and a published 1080p rate. The trade-off is token-based billing, which takes a calculation to turn into cost per second.
- Replicate is easy to budget: one price per second of output, with the lowest published 480p/720p rates among resellers.
- OpenRouter suits teams already using it for text models, because the API is OpenAI-compatible and you only change the model slug.
- fal costs more per second but documents its token formula, endpoints (text-to-video, image-to-video, reference-to-video) and worked examples, such as "30s at 720p 16:9" for "~$13.87."
- Atlas Cloud publishes one "from $0.134" per-second rate across its three endpoints, which makes rough forecasting simple.
- WaveSpeedAI lists the widest endpoint set, including turbo, video-extend, video-edit and motion-control, priced per output.
Setup steps for that official path, including the model ID and how token billing is calculated, are in the BytePlus Seedance 2.5 API guide.
Seedance API Providers FAQ
Which Seedance 2.5 API provider is the cheapest right now?
For Seedance 2.5 Lite, the Seedance 2.5 API at /docs/api: 10 credits per output second, about $0.035–$0.05 per second on annual plans. Among third-party resellers, Replicate lists the lowest rates at $0.1028 (480p) and $0.2312 (720p) per second.
What is the best Seedance 2.5 API?
It depends on your stack. The Seedance 2.5 API at /docs/api gives free quotes, a price cap and the same credits as the web generator; BytePlus ModelArk offers official token billing; OpenRouter suits teams that want one key for many models. Most users integrate on their own by following the /docs/api documentation; if you need more permissions or have special requirements, contact support.
How long can a Seedance 2.5 API video be?
Up to 30 seconds. The Seedance 2.5 API docs list whole-second durations from 4 to 30 seconds for Seedance 2.5, and 4 to 15 seconds for Seedance 2.5 Lite.
Can I get 1080p from the Seedance 2.5 API?
Yes. Seedance 2.5 supports 480p, 720p and 1080p in the Seedance 2.5 API, and BytePlus ModelArk publishes a 1080p token rate. Replicate, OpenRouter, Atlas Cloud and WaveSpeedAI do not list a 1080p price on their official pages.
Can I try Seedance 2.5 for free before paying for an API?
Yes. You get free credits on sign-up to test Seedance 2.5 in the web generator before choosing an API plan; API generation at /docs/api uses a paid subscription or purchased credits.
Try Seedance 2.5 Today
Ready to experience the next generation of AI video creation? Try Seedance 2.5 for free and see the difference for yourself.
