This study evaluates 16 Large Language Models (LLMs) on their ability to generate correct Python code for complex LoRaWAN engineering applications under zero-shot conditions.

• Top performers: DeepSeek-V3, GPT-4, Phi-4, and LLaMA-3.3 demonstrated superior capability in generating accurate code
• Engineering tasks: Models tackled drone placement optimization and power calculation problems of increasing complexity
• Lightweight models: Research shows promising results for locally-executable LLMs, potentially enabling on-device engineering solutions
• Zero-shot effectiveness: Leading models successfully handled natural language prompts without specialized training

For engineering applications, these findings suggest that advanced LLMs can automate complex technical tasks, reducing development time and enabling efficient solutions for drone-based communications infrastructure.

DeepSeek-V3, GPT-4, Phi-4, and LLaMA-3.3 generate correct code for LoRaWAN-related engineering tasks