Cursor Launches 2.0 Update with Composer Model and Multi-Agent Interface

Cursor has released version 2.0, introducing its first coding model, Composer, and a redesigned interface for managing multiple agents in parallel, announced on the company’s blog.

Composer is a frontier model designed for low-latency coding, completing most interactions in under 30 seconds. It uses codebase-wide semantic search to better understand and operate within large codebases. Early testers found it effective for iterative, multi-step coding tasks.

The new interface centers around agents rather than files, allowing users to focus on outcomes while agents handle implementation details. Cursor 2.0 supports running multiple agents simultaneously, using git worktrees or remote machines to prevent interference. Users can also compare outputs from different models to select the best result.

Additional improvements include streamlined tools for reviewing agent-generated code and a built-in browser feature that enables automated testing and iteration. Cursor 2.0 is now available for download on the company’s website.