Alibaba Cloud: Wan 3.0 Prime

Wan 3.0 Prime is the speed-optimized version of the Wan 3.0 video model. It is useful for iterating on reference-driven scenes and frame transitions when several takes need to be explored in a production workflow.

Compare models

wan-3.0-prime

AvailableAlibaba CloudVideoAsync

Price: From $0.14
Modalities: Video

About Wan 3.0 Prime

Wan 3.0 Prime is the speed-focused version of Alibaba's Wan 3.0 video model, listed by Alibaba as wan3.0-video-prime. Its inputs and outputs match the standard model: text, reference images, video and audio, first and last frames, editing, and extension. The difference is turnaround, which makes it the practical choice when many takes of the same scene must be compared.

Where it works well

Alibaba calls it the high-speed version with capabilities aligned to the standard model, so it keeps the same reference, frame, and editing inputs.
Faster generation shortens each review cycle when exploring several variants of a scene.
It accepts the same long clip lengths, up to thirty seconds, as Wan 3.0.
Reference images, videos, and audio can anchor the cast and look across repeated takes.

When to choose another model

Both versions offer the same features, but a final delivery render may still be worth running on standard Wan 3.0 for comparison.
Like the standard model, it is an input-heavy interface, and plain prompts are easier to run on Wan 2.6.

Getting started

Create API key
Create a key in Console, then use it with every model on the platform.

Send your first request

Copy the example for your language and run it against the endpoint.

Image to videoReference to videoStart and end frames to videoText to videoTokenLab endpoint

POST/v1/videos/generations

curl -X POST "https://api.tokenlab.sh/v1/videos/generations" \
  -H "Authorization: Bearer sk-xxx" \
  -H "Content-Type: application/json" \
  -d '{
  "aspect_ratio": "16:9",
  "duration": 5,
  "operation": "text-to-video",
  "resolution": "720p",
  "model": "wan-3.0-prime",
  "prompt": "Cinematic sunrise over a calm lake with gentle camera motion."
}'

Parameters

Image to video

Parameter	Values and limits	Default	Notes
`aspect_ratio`	1:1, 3:4, 4:3, 9:16, 16:9, 21:9	Not provided	Not provided
`duration`	2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30	Not provided	2-30s. The omitted-field default is not verified.
`image`	Not provided	Not provided	The current public contract does not specify an input count limit for this field. Do not infer a count from the example request.
`image_url`	Not provided	Not provided	The current public contract does not specify an input count limit for this field. Do not infer a count from the example request.
`operation`Required	Not provided	Not provided	Not provided
`prompt`	Not provided	Not provided	Not provided
`resolution`	480p, 720p, 1080p	Not provided	Not provided
`start_image`	Not provided	Not provided	The current public contract does not specify an input count limit for this field. Do not infer a count from the example request.

Reference to video

Parameter	Values and limits	Default	Notes
`aspect_ratio`	1:1, 3:4, 4:3, 9:16, 16:9, 21:9	Not provided	Not provided
`audio_url`	Not provided	Not provided	The current public contract does not specify an input count limit for this field. Do not infer a count from the example request.
`audio_urls`	Not provided	Not provided	The current public contract does not specify an input count limit for this field. Do not infer a count from the example request.
`duration`	2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30	Not provided	2-30s. The omitted-field default is not verified.
`image_urls`	Not provided	Not provided	The current public contract does not specify an input count limit for this field. Do not infer a count from the example request.
`operation`Required	Not provided	Not provided	Not provided
`prompt`	Not provided	Not provided	Not provided
`reference_images`	Not provided	Not provided	The current public contract does not specify an input count limit for this field. Do not infer a count from the example request.
`resolution`	480p, 720p, 1080p	Not provided	Not provided
`video_url`	Not provided	Not provided	The current public contract does not specify an input count limit for this field. Do not infer a count from the example request.
`video_urls`	Not provided	Not provided	The current public contract does not specify an input count limit for this field. Do not infer a count from the example request.

Official Wan 3.0 supports image, video, and audio reference inputs for reference-based generation. With reference videos, the billed seconds are the reference video seconds (at most 15 s in total) plus the output seconds; image and audio references are not billed.

Start and end frames to video

Parameter	Values and limits	Default	Notes
`aspect_ratio`	1:1, 3:4, 4:3, 9:16, 16:9, 21:9	Not provided	Not provided
`duration`	2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30	Not provided	2-30s. The omitted-field default is not verified.
`end_image`	Not provided	Not provided	The current public contract does not specify an input count limit for this field. Do not infer a count from the example request.
`image`	Not provided	Not provided	The current public contract does not specify an input count limit for this field. Do not infer a count from the example request.
`image_url`	Not provided	Not provided	The current public contract does not specify an input count limit for this field. Do not infer a count from the example request.
`operation`Required	Not provided	Not provided	Not provided
`prompt`	Not provided	Not provided	Not provided
`resolution`	480p, 720p, 1080p	Not provided	Not provided
`start_image`	Not provided	Not provided	The current public contract does not specify an input count limit for this field. Do not infer a count from the example request.

Official Wan 3.0 supports first and last frame generation; the first/last-frame mode does not accept driving audio.

Text to video

Parameter	Values and limits	Default	Notes
`aspect_ratio`	1:1, 3:4, 4:3, 9:16, 16:9, 21:9	Not provided	Not provided
`duration`	2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30	Not provided	2-30s. The omitted-field default is not verified.
`operation`Required	Not provided	Not provided	Not provided
`prompt`	Not provided	Not provided	Not provided
`resolution`	480p, 720p, 1080p	Not provided	Not provided

Use TokenLab create endpoint and follow returned poll_url; if you need a fixed status path, use `/v1/tasks/{id}`.

Wan 3.0 and Wan 3.0 Video Prime support text-to-video, image-to-video, first/last-frame generation, and reference-based generation with image, video, and audio inputs.

Generated clips can include synchronized audio; output_audio cannot turn it off.

Pricing

The Verified price applies to Verified, which costs less on most models. The Official price is the model maker's published price and applies to the more reliable Official route. Auto bills the route that completes the request.

480p

per second

Official price: $0.068
Verified price: —
Discount: —

720p

per second

Official price: $0.14
Verified price: —
Discount: —

1080p

per second

Official price: $0.28
Verified price: —
Discount: —

	Official priceper second	Verified priceper second	Discount
480p	$0.068	—	—
720p	$0.14	—	—
1080p	$0.28	—	—

Usage & activity

Success rate is the share of requests that completed. Latency is how long a full response takes; P95 means 95% of requests finished within that time.

Usage & availability

Last 24 hours

Requests
Success rate
P95 latency
Total tokens

Model performance

Last updated Nov 1, 2026, 11:26 AM UTC

Metrics appear once privacy and data volume thresholds are met.

Data is based on aggregate user requests, excluding status checks.

Open in Console

Open Wan 3.0 Prime in Console with a prompt ready to edit or send.

Help me create with wan-3.0-prime using text-to-video at /v1/videos/generations. Show the result, status, and cost.

Use cases

Best for

Text to Video

Rapid take comparison
Run the same storyboard with different camera instructions and choose the best take before spending time on a final version.
Storyboard animatics
Turn planned frames into moving previews that show pacing and transitions to a client.
Social variant testing
Generate several cuts of one product message with different openings and measure which one holds attention.
Reference tuning
Swap reference images or voice tracks between runs to find the combination that keeps a character consistent.

Prompt examples

Same scene, three takes: a slow dolly-in, a handheld follow, and a high crane shot over the market stalls.

Use this logo animation as the last frame, and build a five-second reveal that starts with scattered paper pieces.

With this reference voice, show the founder presenting the roadmap on stage, cutting to slides every few seconds.

FAQ

Is Wan 3.0 Prime lower quality than Wan 3.0?

Alibaba describes Prime as speed-optimized with capabilities aligned to the standard version, not as a smaller model. Output can still differ run to run, so compare a few of your own prompts before choosing it for delivery work.

When should I pick Prime over standard Wan 3.0?

Pick Prime during iteration, when you are testing prompts, frames, and references and want answers sooner. Keep standard Wan 3.0 for the render you plan to publish, if you find a visible difference on your content.

Does Wan 3.0 Prime take the same inputs?

Yes. Alibaba lists identical input types: text, reference images, reference videos and audio, documents and web pages, plus first and last frames. Editing and extension are included.

How long can a Prime clip be?

Alibaba documents two to thirty seconds for the Wan3.0 pair, or a smart mode that sets the length. Resolutions are 480p, 720p, and 1080p, and audio is generated unless you turn it off.

Is Prime a new generation after Wan 3.0?

No. Both belong to Wan3.0; Prime is a faster variant of the same model family with the same capabilities, so choose between them on turnaround.

How much does Wan 3.0 Prime cost?

On TokenLab, Wan 3.0 Prime costs $0.068-$0.28 per second. The pricing table above shows the full breakdown. Rates depend on the billing unit, specification, and usage. Compare matching conditions in the model's detailed pricing; a single rate does not determine the total cost.

Which endpoint should Wan 3.0 Prime use?

Use https://api.tokenlab.sh/v1/videos/generations for Wan 3.0 Prime. The request example below shows the matching code shape.

Which operations does Wan 3.0 Prime support?

Wan 3.0 Prime supports Image to video, Reference to video, Start and end frames to video, Text to video. Select an operation above to see its endpoint and request example.