Cursor upgrades to better AI models, enabling developers to tackle larger coding projects

Cursor, an AI-powered code editor, announced support for more capable AI models that can understand and work with larger codebases. The upgrade allows developers to give the AI bigger chunks of code to analyze and modify at once, compared to previous limitations.

Unlike basic autocomplete tools, the improved models can reason through multi-step coding problems, spot errors in their own suggestions, and fix them — similar to how a developer reviews their own code. This means less back-and-forth correction between human and AI.

For software developers and engineering teams, this changes what's practical to automate: instead of small helper tasks (completing one line of code), developers can now hand off larger features or refactoring work to Cursor and iterate once, saving hours of manual coding per day.