Wan 3.0 API (Wan3.0) on APIMart gives you Alibaba's newest Wan video model through the Wan API. Send a prompt alone, a first frame with an optional last frame, or image, video, audio, document and webpage references, and get a 2–30 second video at 480P, 720P or 1080P with audio on by default. Model ID: wan3.0-video.
Generate from a prompt alone or add image, video, and audio references
General / reference lane
Use a prompt alone or add references, documents, or webpages
Frame lane
The first frame is required and the last frame is optional
Frame inputs cannot be mixed with reference images, videos, audio, files, or webpages in one request.
Enter a prompt or add at least one reference image, video, or audio item.
Upload up to 5 WAV/MP3 files, each 1–15 seconds and 15 MB; total duration up to 15 seconds.
WAV, MP3 (max 15MB, up to 5 audios, duration ≤ 15s)
0/5 audio items added. Uploaded files and URLs share the same limit.
Choose auto duration (-1) or an integer from 2 to 30 seconds. With reference video, input plus output duration cannot exceed 30 seconds.
1080P is the default and highest-priced resolution tier.
Generate an audio track by default. Turning it off does not change the price.
Add the provider watermark to the generated video.
Stored on the APIMart CDN
Transparent pricing with no hidden fees. Pay only for what you use.
| Variant | Tier | GoldCurrent price | Platinum | Diamond | Official | Your savings |
|---|---|---|---|---|---|---|
wan3.0-video | 480P | 0.3288Credits(~$0.0329)/sec | 0.3124Credits(~$0.0312)/sec | 0.2959Credits(~$0.0296)/sec | 0.411Credits(~$0.0411)/sec | 20% |
| 720P | 0.6575Credits(~$0.0658)/sec | 0.6246Credits(~$0.0625)/sec | 0.5918Credits(~$0.0592)/sec | 0.8219Credits(~$0.0822)/sec | 20% | |
| 1080P | 1.315Credits(~$0.1315)/sec | 1.2493Credits(~$0.1249)/sec | 1.1835Credits(~$0.1184)/sec | 1.6438Credits(~$0.1644)/sec | 20% |
* Actual costs are subject to final output.
* Total cost = (reference video duration + generated video duration) × the selected model and resolution’s original per-second price × its actual payable percentage ÷ 100.
Wan API access to Alibaba Wan 3.0 on APIMart: turn a prompt, first and last frames, or image, video, audio and document references into a 2–30 second video at 480P, 720P or 1080P, with audio on by default and per-second billing.
50K+
Active Users
99.9%
Uptime
2x
Faster
—
Cost Savings
Call the Wan API with one POST request. Generation is asynchronous: submit a task, then poll the task endpoint for the video URL.
Save your APIMart key in the APIMART_API_KEY environment variable and send it as Authorization: Bearer <key> on every Wan API request.
POST /v1/videos/generations with model wan3.0-video, your prompt, duration, resolution and size. Add two image_urls for first and last frames. The task ID is in data[0].task_id.
Poll GET /v1/tasks/{task_id} every 5–10 seconds until status is completed, then read the video URL from data.result.videos[0].url[0] and save the file to your own storage.
See the Wan 3.0 API reference for every parameter, input rule and response field.
curl -X POST https://api.apimart.ai/v1/videos/generations \
-H "Authorization: Bearer $APIMART_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "wan3.0-video",
"prompt": "A paper boat drifts down a rain-soaked city street at dusk, neon reflections on the water, slow tracking shot, soft rain sounds",
"duration": 5,
"resolution": "720P",
"size": "16:9"
}'import os
import time
import requests
API_BASE = "https://api.apimart.ai"
headers = {"Authorization": f"Bearer {os.environ['APIMART_API_KEY']}"}
# 1. Submit the generation task
payload = {
"model": "wan3.0-video",
"prompt": "A paper boat drifts down a rain-soaked city street at dusk, neon reflections on the water, slow tracking shot, soft rain sounds",
"duration": 5,
"resolution": "720P",
"size": "16:9"
}
resp = requests.post(f"{API_BASE}/v1/videos/generations", headers=headers, json=payload, timeout=60)
resp.raise_for_status()
task_id = resp.json()["data"][0]["task_id"]
# 2. Poll the task until it finishes
for _ in range(200):
task_resp = requests.get(f"{API_BASE}/v1/tasks/{task_id}", headers=headers, timeout=30)
task_resp.raise_for_status()
task = task_resp.json()["data"]
if task["status"] == "completed":
print(task["result"]["videos"][0]["url"][0])
break
if task["status"] == "failed":
raise RuntimeError(task.get("error", {}).get("message", "generation failed"))
time.sleep(10)
else:
raise TimeoutError(f"task {task_id} did not finish in time")curl -X POST https://api.apimart.ai/v1/videos/generations \
-H "Authorization: Bearer $APIMART_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "wan3.0-video",
"prompt": "The camera glides from the first frame to the last frame, morning light slowly fills the room",
"image_urls": [
"https://example.com/frame-start.jpg",
"https://example.com/frame-end.jpg"
],
"duration": 5,
"resolution": "720P"
}'TASK_ID="task_xxx" # replace with the task_id returned by the POST request
curl "https://api.apimart.ai/v1/tasks/$TASK_ID" \
-H "Authorization: Bearer $APIMART_API_KEY"| Parameter | Type | Required | Default | Description |
|---|---|---|---|---|
model | string | Required | Default: wan3.0-video | wan3.0-video, or wan3.0-video-prime with the same parameters. |
prompt | string | Required | Default: — | Scene, motion, camera and sound description, up to 20,000 characters. Optional when you send image, video, audio, file or link inputs. |
duration | integer | Optional | Default: 5 | Length in seconds, 2 to 30, or -1 to let the model choose. With a reference video, input plus output must stay within 30 seconds. |
resolution | string | Optional | Default: 1080P | 480P, 720P or 1080P. Price per second rises with resolution. |
size | string | Optional | Default: adaptive | adaptive, 16:9, 4:3, 1:1, 3:4 or 9:16. |
audio | boolean | Optional | Default: true | Generate a soundtrack. Does not change the price. |
image_urls | array | Optional | Default: — | Public image URLs. Without other media, the first is the first frame and the second the last frame; with video, audio, file or link inputs they become reference images (up to 10). |
image_with_roles | array | Optional | Default: — | Images with an explicit role: first_frame, last_frame or reference_image. |
generation_type | string | Optional | Default: — | frame or reference. Decides how bare image_urls are used. |
video_urls | array | Optional | Default: — | Up to 5 reference clips (mp4 or mov), up to 15 seconds in total. |
audio_urls | array | Optional | Default: — | Up to 5 reference audio clips (wav or mp3), up to 15 seconds in total. audio_url accepts a single clip. |
file_url / link_url | string | Optional | Default: — | One document (up to 100 MB and 50 pages) or one public webpage. Use one or the other, not both. |
seed | integer | Optional | Default: — | 0 to 2147483647 for more repeatable results. |
watermark | boolean | Optional | Default: false | Add a watermark to the output. |
nsfw_check | boolean | Optional | Default: false | Moderate the prompt and input images with omni-moderation-latest before submission. |
Compare every Alibaba Wan video model on APIMart. One API key calls all of them, billed per second of video.
| Model | Best for | Resolution | Duration |
|---|---|---|---|
| Wan 3.0 APIThis page | Longest Wan clips, with first and last frames or image, video, audio, document and webpage references | 480P / 720P / 1080P | 2–30 s |
| Wan 3.0 Prime API | Same parameters as Wan 3.0 under a separate model and price | 480P / 720P / 1080P | 2–30 s |
| Wan 2.7 API | Text to video, first and last frames, video continuation and audio-driven clips | 720P / 1080P | 2–15 s |
| Wan 2.7 R2V API | Keeping characters consistent from reference images or videos | 720P / 1080P | 2–15 s |
| Wan 2.7 Video Edit API | Editing or restyling an existing video with an instruction | 720P / 1080P | Follows source, up to 10 s |
| Wan 2.6 API | Multi-shot text to video and image to video with audio and effect templates | 720p / 1080p | 5 / 10 / 15 s |
| Wan 2.6 I2V API | Dedicated image to video from a single first frame | 720p / 1080p | 5 / 10 / 15 s |
| Wan 2.6 I2V Flash API | Image to video at any length from 2 to 15 seconds, with or without audio | 720p / 1080p | 2–15 s |
| Wan 2.5 API | Short clips that always include sound, with a low-cost 480p tier | 480p / 720p / 1080p | 5 / 10 s |
What the wan3.0-video model accepts, what it returns and the rules that keep requests valid
Workflows that use the length and inputs only Wan 3.0 offers
From sign-up to your first Wan 3.0 video in three steps
Register on APIMart, add balance and create an API key in the dashboard. The same key works for every Wan model.
Try prompts, frames, references, durations and resolutions above and check the cost estimate before you write code.
Submit tasks from your backend with the example below, then poll for the result.
Answers about Wan 3.0, the Wan API on APIMart and how to choose between Wan models
Wan 3.0 is Alibaba's newest Wan video model. It creates 2–30 second videos at up to 1080P from text, keyframes or image, video, audio, document and webpage references, with audio included by default. On APIMart its model ID is wan3.0-video.
Alibaba opened the Wan 3.0 public beta in August 2026, with API access through Alibaba Cloud Model Studio. The wan3.0-video model is available through the APIMart Wan API now.
The Wan API on APIMart is one asynchronous endpoint, POST /v1/videos/generations, for every Wan video model: wan3.0-video, wan3.0-video-prime, wan2.7, wan2.6 and wan2.5-preview, plus the Wan 2.7 and 2.6 variants. You submit a task, receive a task ID and poll for the finished video, all with one APIMart key.
Use Wan 3.0 for clips up to 30 seconds or when you need audio, video, document or webpage references. Use Wan 2.7 for 2–15 second clips with first-last-frame control or video continuation, Wan 2.6 for multi-shot clips and effect templates, and Wan 2.5 Preview for 5 or 10 second clips with sound. The comparison table on this page links to every Wan model.
Wan 3.0 doubles the maximum length to 30 seconds, adds a 480P tier, generates audio by default and accepts reference images, videos and audio clips plus document and webpage inputs. Wan 2.7 stays at 2–15 seconds and 720P or 1080P, but offers negative prompts, continuation of a source clip and driving audio for image to video.
wan3.0-video and wan3.0-video-prime accept exactly the same parameters, so switching only changes the model field. They are priced separately; compare them on the Wan 3.0 Prime page before you choose.
model is required, and prompt is required unless you send media. Optional fields are duration (2–30 or -1, default 5), resolution (480P, 720P or 1080P, default 1080P), size (adaptive, 16:9, 4:3, 1:1, 3:4 or 9:16), audio (default true), seed, watermark, nsfw_check, generation_type, image_urls, image_with_roles, video_urls, audio_urls or audio_url, and file_url or link_url.
Yes. Send one image in image_urls for a first frame, or two images for first and last frames; you can also label them with image_with_roles. With only images and no other media, the first image is treated as the first frame and the second as the last frame.
No. Frame inputs (first_frame, last_frame) and reference inputs (reference images, videos, audio, file_url and link_url) cannot be combined. A mixed request returns HTTP 400 before a task is created and is not charged. Use generation_type to say whether bare image_urls are frames or references.
Generation usually takes a few minutes, and longer or higher-resolution videos take longer. Poll GET /v1/tasks/{task_id} every 5–10 seconds until status is completed, then read the video URL from data.result.videos[0].url[0] and save it to your own storage.
Wan 3.0 is billed per second of video, and the rate depends on the resolution: 480P costs least and 1080P, the default, costs most. Audio does not change the price, and with duration -1 you pay for the actual output length. The pricing table and its notes on this page show current rates; Gold members save 20%, and Platinum and Diamond members save up to 28%.
No. APIMart has no free quota for Wan 3.0; you pay as you go for each successful task. Alibaba has not published Wan 3.0 weights, so it is available as a hosted API rather than a download. Alibaba's open-weight Wan releases are the Wan 2.2 family and specialized models such as Wan-Dancer; Wan 2.5 and later general video models are API-only.
Sign up on APIMart, add balance and create a key on the API Keys page. Send it as Authorization: Bearer <key>. The same key calls Wan 3.0, Wan 2.7, Wan 2.6, Wan 2.5 and every other model on APIMart.
Yes. Wan3.0, Wan 3, Wan3 API and Alibaba Wan 3.0 all refer to the same model, which APIMart exposes as wan3.0-video. Wan 3.0 Prime (wan3.0-video-prime) is a separate model with its own page.
Explore more models in the same category.

Seedance 2.5
Seedance 2.5 is ByteDance's audio-video model for text-to-video, multimodal reference, editing, extension and first-last-frame control. Generate up to 30 seconds in one job with up to 30 images, 10 videos and 10 audio clips as references, at up to 1080p.

Flux 3 Video
FLUX 3 Video by Black Forest Labs supports text-to-video, image-to-video with ordered keyframes and video continuation. It generates 5–20 second HD or FHD clips with synchronized audio in seven aspect ratios.

MiniMax H3
MiniMax H3 is MiniMax's video model for text-to-video, image-to-video with first and last frames, and multimodal reference-to-video using images, videos and audio. It outputs 4–15 second clips at 2K or 768P with an audio track.

Gemini Omni Flash Preview
gemini-omni-flash-preview is a multimodal video generation and editing model launched by Google.