Skip to main content

​
Short Answer

The default group is already the recommended route for normal image generation and is generally fast enough. The enterprise group is primarily a fallback when the default route has an incident; it is not a dedicated acceleration route for reducing image-generation time.

​
Why the Enterprise Group May Not Be Faster

APIYI already uses optimized return routes across China Telecom, China Unicom, and China Mobile. The main source of image-generation latency is usually upstream model inference rather than APIYI network routing. Switching to the enterprise group therefore generally cannot reduce the model’s native generation time.
The main difference between groups is routing and failover, not faster model inference.
  • Default group: Recommended for normal requests, with optimized speed and stability
  • Enterprise group: Used as a fallback during default-route incidents, with a focus on availability

​
What If Generation Speed Is the Priority?

If your workload prioritizes faster image output, test the Nano Banana image model, then decide based on image quality, cost, and measured generation time.
Model speed varies with image dimensions, output count, prompt complexity, and upstream load. No route or group can guarantee a fixed image-generation time. Test with your real request parameters and production-like concurrency.