Both at 100 credits. WAN offers 15-second clips while Vidu adds direct audio and motion control.
Vidu Q3 from Shengshu and WAN v2.7 from Alibaba represent two different approaches to AI video generation. Shengshu's latest video model with native audio, reference-to-video for character consistency, and up to 1080p resolution. On the other side, alibaba's latest video model with 1080p output, reference-to-video, and up to 15-second clips.
Both models cost 100 credits per generation, so the decision comes down to capabilities rather than budget. Vidu Q3 excels at character-consistent video, reference-based generation, video with audio, while WAN v2.7 is better suited for multi-shot storytelling, reference-based generation, longer 1080p clips.
Updated August 2026
Both cost 100 credits, so pick based on your needs. Go with Vidu Q3 for character-consistent video. Choose WAN v2.7 if you need multi-shot storytelling. Since they're the same price on Melies, try both and see which output style you prefer.
| Provider | Shengshu | Alibaba |
| Released | May 2026 | Jul 2026 |
| Cost | 100 | 100 |
| Speed | Medium | Medium |
| Max duration | 10 seconds | 15 seconds |
| Image input | ||
| Audio | ||
| Resolutions | 540p, 720p, 1080p | 720p, 1080p |
Both models cost the same. Here's how credits add up at scale.
| Volume | Vidu Q3 | WAN v2.7 |
|---|---|---|
| 5 videos | 500 | 500 |
| 10 videos | 1,000 | 1,000 |
| 25 videos | 2,500 | 2,500 |
Vidu Q3 generates native audio alongside the video — no need to add sound in post-production.
WAN v2.7 supports up to 15 second clips, giving you more room for complete scenes.
At 100 credits per video, Vidu Q3 lets you generate more clips and pick the best ones.
Practical answers before you start creating.





Compare Vidu Q3, WAN v2.7, and 10+ AI video models in one workspace. Switch models freely, same credits.
Start Creating Videos