Pre-checks
What problem does this solve?
Solve the needs to test multiple models to determine the model needed for an agent execution.
Proposed solution
Having a LileLLM containers in the same private Network as Clawith (or integrated Clawith multi-models management) and give models self A/B testing feature to determine after a certain period of time which model suits best its needs without being overkill.
Willing to contribute?