Human-Level Reasoning: A Comparative Study of Large Language Models on Logical and Abstract Reasoning