This research provides a systematic approach to evaluate Large Language Models in code generation tasks, addressing the growing demand for automated software development tools.

• Establishes standardized metrics for assessing code quality, functionality, and efficiency
• Reviews the historical evolution of LLMs for code generation
• Identifies current evaluation challenges and methodological gaps
• Proposes improved evaluation frameworks for engineering applications

For engineering teams, this research offers valuable insights into selecting appropriate LLMs for development workflows and understanding their capabilities and limitations in real-world coding scenarios.

A Survey on Evaluating Large Language Models in Code Generation Tasks