What it measures
- CursorBench 4.0
CursorBench 4.0
Version 4.0 coding-agent evaluation; distinct from CursorBench 3.2.0.
How to read the score
% score; higher is better.
Sourced Ranking
| Rank | Model | Provider | Score | Source | Source Type | Verified |
|---|---|---|---|---|---|---|
| 1 | Claude Opus 5.5 | Anthropic | 57.8 | Introducing Claude Opus 5.5 Accessed 2026-10-02 | official | Yes |
| 2 | Claude Sonnet 5.5 | Anthropic | 55.5 | Introducing Claude Sonnet 5.5 Accessed 2026-10-02 | official | Yes |
Limitations
- Scores may not be directly comparable across model reports, harnesses, or task releases.
- Preserve the exact version, evaluation configuration, and reporting source in each result row.
- Vendor and third-party scores are stored as raw evidence and are not automatically used in the five-axis capability radar.
leaderboard
CursorBench 4.0 source record
Preserve the exact version, evaluation configuration, and reporting source in each result row. Vendor and third-party scores are stored as raw evidence and are not automatically used in the five-axis capability radar.