[Feature] Add LLM Test interface

[666haiwen]

What problem does this feature solve?

Add a new large model testing interface to ensure the connectivity of the large model, or provide clear error feedback in case of disconnection.

What does the proposed API look like?

VMind.llmTest