Define a set of models that Jupyter AI should be tested with when adding new functionality

[andrii-i]

trafficstars

Problem

Set of models that Jupyter AI supports / provides is continiously increasing in size. As a result, testing new functionality or changes with all models is not feasible and will become even less feasible the more models are supported / provided.

Proposed Solution

Define a set of models that Jupyter AI should be tested with when adding new functionality or making changes.