## PCEval: Evaluating LLM Capabilities in the Physical World Large Language Models (LLMs) have demonstrated remarkable capabilities across various domains, including software development. However, their effectiveness has not been fully explored when hardware constraints are considered, such as in physical computing, where software must interact with and control physical hardware. To address this gap, extsc{PCEval} (Physical Computing Evaluation) has been introduced, the first benchmark in physical computing that enables a fully automatic evaluation of the capabilities of LLMs in both the logical and physical aspects of the projects, without requiring human assessment. The evaluation framework assesses LLMs in generating circuits and producing compatible code across varying levels of project complexity. Findings from comprehensive testing of 13 leading models reveal that while LLMs perform well in code generation and logical circuit design, they struggle significantly with physical breadboard layout creation, particularly in managing proper pin connections and avoiding circuit errors. extsc{PCEval} advances our understanding of AI assistance in hardware-dependent computing environments and establishes a foundation for developing more effective tools to support physical computing education. Benchmarks are fundamental for measuring and comparing the performance of artificial intelligence systems. They allow for the identification of strengths and weaknesses, guiding the development of more effective and performant solutions. In the context of physical computing, a benchmark like PCEval can help improve the integration between software and hardware, opening new possibilities for the automation and control of physical systems.