From a functional perspective I think LLM's problem solving abilities are something I would call a type of reasoning/cognition. If you want to call LLM's ability to do this by a different name from when humans do it, that's fine I suppose, but I don't think that is a very helpful definition.
