Wan 2.5
What is Wan 2.5?
Wan 2.5 is the audio milestone in Alibaba's Wan video line. Two modes are exposed on Unifically: text-to-video from a prompt, and image-to-video from a start_image_url. Output runs at 480p, 720p, or 1080p in 5 or 10 second clips. Audio is auto-generated by default, can be replaced with a custom track via audio_url for hard-synced soundtracks, or turned off entirely for silent output. Multi-shot and reference-to-video are not on this model. Use it when you want short clips with synchronized audio without jumping to the multi-shot complexity that arrives in Wan 2.6.
Key features of Wan 2.5
Four features cover what changes versus Wan 2.2.
Auto audio in the same call
Audio generation is on by default. The model returns a clip with a synchronized audio track without a separate text-to-audio pipeline.
Custom audio sync via audio_url
Upload a wav/mp3 (max 15MB, 3–30 seconds) and the model syncs visuals to your track. Useful when the audio mix is fixed and visuals need to land on it.
5 or 10 second clips at 480p, 720p, or 1080p
Wider duration and resolution range than Wan 2.2. Useful for short-form social and paid-ad slots that need a 1080p master.
Negative prompts and seeds
Suppress unwanted content with `negative_prompt` and pin the result with a seed. Same brand-safe controls as earlier Wan video models.
Best for
Short ads with audio
5 or 10 second cuts with optional generated or custom audio in one call.
Social posts at three resolutions
Square, vertical, or wide formats across 480p, 720p, and 1080p.
Explainer beats with sound design
Quick motion from an illustration or product still with synchronized audio.
Custom audio sync
Upload your own track when the audio mix is fixed and visuals must hit specific beats.
Negative-prompt control
Suppress unwanted content for brand-safe output.
Prototype reels
Test hooks and cuts before moving to longer Wan 2.6 multi-shot clips.
API examples
Call Wan 2.5 from any language by POSTing to /v1/tasks. Full parameter docs live at docs.unifically.com/models/video/alibaba/wan-2.5-video.
curl -X POST https://api.unifically.com/v1/tasks \
-H "Content-Type: application/json" \
-H "Authorization: Bearer YOUR_API_KEY" \
-d '{
"model": "alibaba/wan-2.5-video",
"input": {
"prompt": "Sunset over ocean waves",
"mode": "t2v",
"resolution": "1080p",
"duration": 10
}
}'
Successful submission returns a task_id. Poll GET /v1/tasks/<task_id> or set a callback_url on the request to receive the finished result.
FAQs
People also ask
Wan 2.5 is the audio-capable Wan video model from Alibaba. It supports text-to-video and image-to-video at 480p, 720p, or 1080p with 5 or 10 second clips. Audio is generated automatically by default; you can also upload a custom audio file for sync or disable audio for silent output.
480p, 720p, and 1080p across both text-to-video and image-to-video.
5 or 10 seconds. The default is 5.
Yes. Pass an audio_url to drive audio-video sync to a fixed track. Or let the model generate audio with audio: true (default), or set audio: false for silent output.
Move to Wan 2.6 when you need clips up to 15 seconds, multi-shot mode (2–8 segments), or reference-to-video for character consistency. Move to Wan 2.7 when you need last-frame, video continuation, or instruction-based video editing.
No. video_urls and multi_prompt are not supported on Wan 2.5. For reference-to-video or multi-shot, use Wan 2.6 or later.
