In e-commerce, accurately capturing the semantic intent of user queries is of high importance for retrieving relevant products. Traditional search engines that rely on keyword matching often fail to grasp the contextual and semantic relationships in user inputs, leading to suboptimal search results. This chapter investigates the application of semantic search methods using advanced language models on a dataset of Turkish e-commerce products. Specifically, the performance of Turkish BERT, LLaMA 3.1-8B, and Voyage AI’s Multilingual-2 models is benchmarked in extracting semantic content and improving search accuracy. The research demonstrates that these models, particularly Voyage AI’s Multilingual-2, significantly outperform traditional methods by leveraging dense vector representations to capture the meaning of user queries, thus enhancing the precision and relevance of search results. The study contributes to the growing body of research on semantic search in diverse linguistic contexts and provides practical implications for improving e-commerce search functionalities.