Cursor releases CursorBench 3.1, an open benchmark for AI code completion

Cursor published an updated evaluation framework for measuring AI-assisted coding performance, now at version 3.1. The benchmark lets developers test how well AI handles real coding tasks, giving the community a concrete way to compare tools beyond marketing claims.