This research evaluates the energy efficiency and performance of code generated by large language models across multiple programming languages and platforms.

Key findings:

• LLM-generated code tends to be less energy efficient than human-written code
• Performance varies significantly across different programming languages
• Models struggle with creating optimized solutions for computationally intensive tasks
• Understanding these efficiency gaps is crucial for sustainable software development

For engineering teams, this research highlights important considerations when integrating AI code generation into development workflows, particularly for performance-critical applications where energy consumption matters.

AI-Powered, But Power-Hungry? Energy Efficiency of LLM-Generated Code