跳转至

cursor-evals-benchmark-3-1-2026-07

Ch01.340 cursor-evals-benchmark-3-1-2026-07

📊 Level ⭐ | 0.7KB | entities/cursor-evals-benchmark-3-1-2026-07.md

-> 原文存档

CursorBench 3.1 is Cursor's coding agent benchmark suite. It evaluates AI agents on ambiguous, multi-file tasks drawn from real Cursor development sessions. Higher scores are better.

来源