Will the LLM be improved with newer AI models as they evolve?

Currently the AI LLM space is evolving rapidly with newer iterations as fast as weekly builds and monthly improvement cycles.

I propose a change request that if Bentley have a permanent AI dev team that are constantly training the models to improve them then perhaps allowing end users to download/installed either at runtime via KeyIn/Command, or a Startup Command Line Switch.

A runtime command line would be like a hot swap where the user can run a command that downloads and implements a newer AI model. A Command Line switch would download and install it at application startup.

Or there could be a Configuration Variable that, when made active, automatically checks for newer AI LLM builds and shows a splash screen for the user to download and install, or offer a postponement for a day or week or silent load.