ALIBABA
Wan2.7 Video
Мультимодальная видео-модель Alibaba: T2V, I2V, референсные изображения, аудио
Strengths
What it's the best tool for
- Thinking Mode for coherent scenes
- Natural-language video editing
- Lip-sync with voice preservation
- 2–15s clips at 720p/1080p
- Text2Video and image2video unified
Limitations
When to reach for something else
- Chinese-policy content filters
- Complex scenes trail Veo 3.1 Pro
- High Thinking adds latency
- No 4K output
Sample output
How Wan2.7 Video responds
Prompt
Replace the character's line in this clip with "Hi, I'm back", preserving voice and facial expression. Leave the rest untouched.
Wan2.7 Video
https://netroom.ai/media/demo/wan-27-video-edit.mp4
Where teams use it
Four scenarios where it pays for itself
01
Localised ads
Dialogue swap with lip-sync
02
Promo clips
Thinking for scripts
03
No-reshoot edits
Natural-language editing
04
Image-born shorts
Image2Video
About model
More about Wan2.7 Video
Wan 2.7 Video — Alibaba's AI Video Suite
Wan 2.7 Video from Alibaba Tongyi (April 2026) is a complete video suite: text2video, image2video, reference video and editing. 2–15 seconds at 720p or 1080p. Hosted on NetRoom.
Capabilities
Native lip-sync with voice preservation when rewriting scripts. Natural-language editing — "change action, scene, camera, outfit". Thinking Mode plans the scene first.
Where it fits
Localised spots, ads with lip-sync, editing finished video without a reshoot.
Try Wan2.7 Video
right now
Free access to basic models. No card, no obligations.