Recent advancements in large language models (LLMs) have shown significant improvements in complex reasoning tasks. Papers have explored various methods to enhance LLM performance, including self-training with AlphaLLM-CPL, structured dialogue, and Monte Carlo Tree Search (MCTS) behavior. The 'Thinking LLMs' paper presents a simple yet effective alternative to OpenAI's o1 model. Another study highlights the benefits of query rephrasing in prompts, boosting in-context learning for zero-shot and few-shot reasoning tasks with EchoPrompt. Additionally, LLMs have demonstrated proficiency in UML modeling, although they still struggle with mapping complex relationships. In a study involving 45 undergraduate students, LLMs aided in UML modeling tasks. Code prompting has also been identified as a method to improve conditional reasoning abilities by transforming natural language problems into code.