The field of large language models is currently moving towards more efficient test-time scaling methods. Researchers are exploring novel frameworks and strategies to enhance the reasoning capabilities of these models while reducing computational overhead. One notable direction is the development of adaptive test-time scaling strategies that dynamically adjust reasoning depth based on question complexity. Another area of focus is the improvement of value-guided search methods, which have shown promising results in achieving better test-time scaling than standard methods. Noteworthy papers include: Value-Guided Search for Efficient Chain-of-Thought Reasoning, which proposes a simple and efficient method for value model training on long-context reasoning traces. T$^2$: An Adaptive Test-Time Scaling Strategy for Contextual Question Answering, which presents a novel framework that dynamically adapts reasoning depth based on question complexity. Stepwise Reasoning Checkpoint Analysis: A Test Time Scaling Method to Enhance LLMs' Reasoning, which introduces checkpoints between reasoning steps to reduce path homogenization and improve reasoning accuracy. First Finish Search: Efficient Test-Time Scaling in Large Language Models, which introduces a training-free parallel decoding strategy that launches $n$ independent samples and returns as soon as any one completes.