One API.
Every route.
The image & video models you want.
The providers, regions and prices behind them.
Set up OpenGen
http://localhost:4310/llms.txtSearch. Compare. Call. One catalog across every interface.
Find your next generation.
All-in-one video generation. Six documented Alibaba regional offerings.
Every regional deployment is its own route.
| Provider / endpoint region | Published price | End-to-end latency | Success rate | |
|---|---|---|---|---|
Beijing, ChinaMainland China alibaba/cn-beijing/wan-3.0 | $0.082513/ output secondPrice source ↗ | —No observations | —0 samples | |
SingaporeInternational alibaba/ap-southeast-1/wan-3.0 | $0.10/ output secondPrice source ↗ | —No observations | —0 samples | |
Tokyo, JapanGlobal inference alibaba/ap-northeast-1/wan-3.0 | $0.082513/ output secondPrice source ↗ | —No observations | —0 samples | |
Frankfurt, GermanyGlobal inference alibaba/eu-central-1/wan-3.0 | $0.082513/ output secondPrice source ↗ | —No observations | —0 samples | |
Virginia, USGlobal inference alibaba/us-east-1/wan-3.0 | $0.082513/ output secondPrice source ↗ | —No observations | —0 samples | |
Hong Kong, ChinaGlobal inference alibaba/cn-hongkong/wan-3.0 | $0.082513/ output secondPrice source ↗ | —No observations | —0 samples |
An endpoint region is not a residency guarantee. Global inference may process requests elsewhere. Success rate measures generation outcomes, not provider uptime.
Catalog listings describe published offerings. Documented does not mean live-verified. Performance stays unknown until real observations exist. Endpoint region and inference location are listed separately.
How we compareA route you can trace.
A bill you can explain.
Same model. Separate routes.
Exact variants, regional deployments and source links stay visible.
Evidence before a score.
No invented speed or reliability. Every measurement needs a sample.
Open from end to end.
Run the catalog and gateway on infrastructure you control.