Inputs
A prompt plus only the image, video, audio, or reference assets explicitly supported by the selected model.
video capability center
Video API: text, image, reference, and editing workflows
Choose Chinese video models by verified task, understand asynchronous jobs, and use a complete create, poll, download, and verify workflow.
Contract
Inputs, outputs, and lifecycle.
Outputs
A task record followed by a downloadable video artifact when the job succeeds.
Endpoint
POST /v1/video/generations; GET /v1/video/generations/{task_id}
Lifecycle
Asynchronous. Create once, preserve the task ID, poll the same job to a terminal state, then download and validate the artifact.
Supported tasks
One definition per task.
text-to-video, image-to-video, reference-to-video, video-editing, video-with-audio
Limit: Duration, resolution, reference counts, audio generation, editing, and accepted asset types vary by model; an unlisted combination is unsupported.
Matching catalog
18 models with explicit metadata.
Unknown capabilities are omitted, never guessed from a model name. Pricing examples use the synchronized public data; live Dash Pricing remains authoritative.
kling-3.0-turbo
Kling 3.0 Turbo — fast generation with audio. 720p 5s with audio ≈ $0.56.
Tasks: text-to-video, image-to-video, video-with-audio
Endpoint: POST /v1/video/generations
Cost example: 1 generation × $0.5556 = $0.5556.
MiniMax-H3
MiniMax H3 (Hailuo 3.0) — omni-modal video with native stereo audio in one pass. 768P/2K, 4-15s. $0.08/s at 768P, $0.13/s at 2K.
Tasks: text-to-video, image-to-video, reference-to-video, video-with-audio
Endpoint: POST /v1/video/generations
Cost example: 1 second × $0.08 = $0.08.
doubao-seedance-2-5-260628
ByteDance Seedance 2.5 — video generation with single-shot output up to 30 seconds, up to 50 multimodal reference assets, and partial editing. 480p/720p output.
Tasks: text-to-video, image-to-video
Endpoint: POST /v1/video/generations
Cost example: The public unit price is not published; use live Pricing before submitting a billable request.
MiniMax-Hailuo-2.3
MiniMax Hailuo 2.3 — text/image-to-video. 768p 6s = $0.28.
Tasks: text-to-video, image-to-video
Endpoint: POST /v1/video/generations
Cost example: 1 generation × $0.2778 = $0.2778.
doubao-seedance-2-0-mini-260615
Seedance 2.0 Mini — lightweight, budget-friendly video generation.
Tasks: text-to-video
Endpoint: POST /v1/video/generations
Cost example: The public unit price is not published; use live Pricing before submitting a billable request.
kling-v3
Kling 3.0 text/image-to-video with 720p/1080p/4K output, 3-15s duration, optional audio, and tiered pricing.
Tasks: text-to-video, image-to-video, video-with-audio
Endpoint: POST /v1/video/generations
Cost example: 1 5-second 720p silent generation × $0.4799 = $0.4799.
kling-v3-omni
Kling 3.0 Omni — reference-based multi-image video generation and editing.
Tasks: reference-to-video, image-editing
Endpoint: POST /v1/video/generations
Cost example: 1 generation × $0.4167 = $0.4167.
MiniMax-Hailuo-02
Hailuo 02 — budget video generation from 512p.
Tasks: text-to-video, image-to-video
Endpoint: POST /v1/video/generations
Cost example: 1 generation × $0.2778 = $0.2778.
doubao-seedance-2-0-260128
ByteDance Seedance 2.0 flagship video model. Billed by output resolution.
Tasks: text-to-video, image-to-video
Endpoint: POST /v1/video/generations
Cost example: The public unit price is not published; use live Pricing before submitting a billable request.
MiniMax-Hailuo-2.3-Fast
Hailuo 2.3 Fast — image-to-video. 768p 6s = $0.19.
Tasks: image-to-video
Endpoint: POST /v1/video/generations
Cost example: 1 generation × $0.1875 = $0.1875.
doubao-seedance-2-0-fast-260128
Seedance 2.0 Fast — cost-efficient video generation. 480p 5s ≈ $0.26.
Tasks: text-to-video, image-to-video
Endpoint: POST /v1/video/generations
Cost example: The public unit price is not published; use live Pricing before submitting a billable request.
happyhorse-1.1-t2v
Alibaba HappyHorse 1.1 (T2V) — text-to-video with improved semantic understanding, camera control, and dynamic generation. Fluid, detailed, physically consistent 720P/1080P video.
Tasks: text-to-video
Endpoint: POST /v1/video/generations
Cost example: 1 second × $0.075 = $0.075.
happyhorse-1.1-r2v
Alibaba HappyHorse 1.1 (R2V) — reference-to-video with up to 9 reference images, strong subject/scene/style consistency and camera control. 720P/1080P.
Tasks: reference-to-video
Endpoint: POST /v1/video/generations
Cost example: 1 second × $0.075 = $0.075.
happyhorse-1.1-i2v
Alibaba HappyHorse 1.1 (I2V) — image-to-video with better texture, ID consistency across clips, motion fluidity, and audio-visual sync. 720P/1080P.
Tasks: image-to-video
Endpoint: POST /v1/video/generations
Cost example: 1 second × $0.075 = $0.075.
wan2.7-t2v
Alibaba Wan2.7 (T2V) — text-to-video with upgraded acting, nuanced emotion, intense action, and dramatic camera cuts. 720P/1080P.
Tasks: text-to-video
Endpoint: POST /v1/video/generations
Cost example: 1 second × $0.1 = $0.1.
wan2.7-r2v
Alibaba Wan2.7 (R2V) — reference-to-video, up to 5 mixed image/video refs + audio-timbre reference, stronger character/prop/scene consistency.
Tasks: reference-to-video
Endpoint: POST /v1/video/generations
Cost example: 1 second × $0.35 = $0.35.
wan2.7-videoedit
Alibaba Wan2.7 VideoEdit — natural-language video editing, local or global, reference-image element replacement, motion/effect/camera replication.
Tasks: No narrower task claim published
Endpoint: POST /v1/video/generations
Cost example: 1 second × $0.2 = $0.2.
wan2.7-i2v
Alibaba Wan2.7 Image-to-Video — first-frame, first/last-frame, and continuation workflows with optional driving audio at 720P or 1080P.
Tasks: image-to-video, video-with-audio
Endpoint: POST /v1/video/generations
Cost example: 1 second × $0.1 = $0.1.
Run or estimate before registration.
Calculate API cost · Generate curl, Python, or JavaScript · Install the Seedance 2 Agent Recipe
Source and method: public model pricing dataset, updated 2026-08-08. Limitations are stated above; third-party claims remain next to the model pages that cite them.