Cursor launches CursorBench 3.1 evaluation suite

Cursor announced the release of CursorBench version 3.1. The suite provides benchmarks for measuring AI‑assisted coding performance. It is hosted on Cursor’s official evaluations page for public access. Developers can use the benchmarks to compare different code‑generation models. The update follows earlier versions that gathered community feedback. CursorBench 3.1 includes revised test cases to reflect current development workflows. The release aims to help researchers gauge model accuracy and efficiency. Cursor encourages users to submit results and contribute to ongoing improvements.