VIDEO API MODELS

Video models. Understand the cost of a finished result.

Compare supported video models by resolution, audio options, duration, and billing unit before you connect a generation workflow.

READ THE UNIT FIRST

Count the seconds that are billed.

All listed video models use per-second pricing. Compare the same resolution, reference-video setting, and audio option to estimate the cost of a finished result.

01 / NO VIDEO REFERENCE

Duration × the matching rate

Without a video reference, multiply output duration by the matching rate. A hypothetical 10-second output at $0.10 per second costs $1 before other charges. This is a calculation example, not a model quote.

02 / WITH VIDEO REFERENCE

Input + output seconds × the rate

For MiniMax H3, Seedance, and Wan, billable duration includes the reference input video plus the generated output. Use the matching reference-video rate. H3 and Wan keep the same rate with or without a reference. Kling VIDEO 3.0 retains output-duration billing; check its audio tiers separately.

FROM REQUEST TO USABLE ASSET

Design for the whole generation job.

  1. 01Validate the request

    Use an enabled model ID, supported settings, and assets you have permission to process.

  2. 02Track the result

    Follow the route's response format. If it returns a task ID, poll the documented status endpoint with backoff.

  3. 03Deliver and reconcile

    Save the result before links expire, review usage, and check task state before retrying a timed-out generation.

SinoRouter uses the gateway https://api.sinorouter.ai/. Endpoint paths, request fields, and status responses vary by enabled model. A working text API does not prove that a video or image route is configured.

Need a route for a larger workload?Send your target resolution, duration, volume, and launch plan to our team for a practical review.info@sinorouter.ai →WhatsApp →