Premium simplicity vs granular frame control — which approach fits your workflow?
WAN v2.2 is no longer offered on Melies. Veo 3.1 is available today. The specs below are kept for reference.
Veo 3.1 from Google and WAN v2.2 from Alibaba represent two different approaches to AI video generation. Google's most advanced video model with native audio, 4K resolution, and reference image support. On the other side, alibaba's image-to-video model with granular frame and interpolation controls.
At 60 credits, WAN v2.2 costs 340 credits less than Veo 3.1 (400 credits). WAN v2.2 is no longer offered on Melies. Veo 3.1 is, along with the other current video models listed on the pricing page.
Updated November 2026
WAN v2.2 is no longer offered on Melies, so Veo 3.1 is the one you can use here. At 400 credits, it is a fit for highest quality video with sound. For other options, compare the current video models and their credit costs on the pricing page.
| Provider | Alibaba | |
| Released | Oct 2025 | Jul 2025 |
| Cost | 400 | 60 |
| Speed | Slower | Medium |
| Max duration | 8 seconds | ~5 seconds at defaults (81 frames / 16fps) |
| Image input | ||
| Audio | ||
| Resolutions | 720p, 1080p, 4K | 480p, 580p, 720p |
WAN v2.2 is 85% cheaper than Veo 3.1. Here's how credits add up at scale.
| Volume | Veo 3.1 | WAN v2.2 |
|---|---|---|
| 5 videos | 2,000 | 300 |
| 10 videos | 4,000 | 600 |
| 25 videos | 10,000 | 1,500 |
Veo 3.1 generates native audio alongside the video — no need to add sound in post-production.
Veo 3.1 is the premium option when every frame counts — final cuts, presentations, or social media covers.
Practical answers before you start creating.





Compare Veo 3.1 with the other video models on Melies in one workspace. Switch models freely, same credits.
Start Creating Videos