Large Language Models (LLMs) can be used by software developers to generate and update code. An LLM-based coding copilot provide code completion or fully generate code from a description. Although coding copilots are a new tool without significant quantitative metrics on their performance, the hope is that they can save time and reduce errors, and allow developers to focus on higher-level tasks.

LLMs are also used for automated debugging, code review, and documentation generation. Additionally, LLMs can be used to generate high-quality documentation for projects.