Large language models in manufacturing: a comprehensive review

Large Language Models (LLMs) have become transformative tools in modern manufacturing. transitioning manufacturing systems from passive data-driven tools toward autonomous decision-making agents capable of reasoning, planning, and supporting complex operational control across the manufacturing lifecycle. This systematic review investigates the application of LLM-based approaches across eight key manufacturing sectors. Examining the studies and datasets that underpin their deployment. Dominant and relevant works were scrutinized to summarize the characteristics of their methodologies, the specifics of implementation, and the principal findings. In addition, limitations and challenges associated with LLM applications were identified and analyzed. Through a root cause analysis, the main sources of these limitations were identified. These insights were then used to develop practical recommendations to improve reliability, interpretability, and integration in industrial settings. This review provides an informative case of LLM use in manufacturing both in a macro-level perspective of the entire value chain and detailed insights that facilitate a deeper understanding of the capabilities, constraints, and opportunities of these models in industrial practice.