CursorBench 4.0

Claude Opus 5.5 57.8% leads since Sep 2026 In use

Ambiguous multi-file coding tasks from real Cursor sessions in Cursor's production harness. Agentic graders score correctness, with cost reported.

Added long-horizon problems: investigation, intent understanding, managing jobs and design adherence. Scores do not compare with 3.x.

Made by Cursor Released Sep 2026 Fixing real repositories Tasks kept private Last in a launch post Sep 2026 Last model added Sep 2026

Level: Expert Specialists with years of training, often over hours of work
Scored: Graded by a model A language model scores each response, usually against a rubric.
Answers: Code A program or patch run against tests.

Leaderboard

Models scored in lab cards

Launch posts list it, from 2 labs

System cards report it, from 2 labs

Claude Opus 5.5 leads with 57.8%

Best score per model as labs report them in their own cards. Labs pick their own settings, so these compare less cleanly than independent runs.

Bars Table

#	Model	Score	Date	Source
1	Claude Opus 5.5	57.8%	Sep 2026	Model card
2	Claude Sonnet 5.5	55.5%	Sep 2026	Model card
3	Grok 4.7	46.3%	Sep 2026	Model card

2 models have held number one since its release in Sep 2026

Lab claims count from the day the card came out, independent runs from the model’s release.

xAI Anthropic

Lab claims 2 leaders

Claude Opus 5.5 57.8%

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Reported by the lab in its model card Best so far

2 labs put it in a launch post in the last six months

Launch posts from every lab that has listed it, one tick per post. Tall colored ticks include it, and short gray ones leave it out.

xAI

Anthropic

Date	Launch post	Score	As the post writes it
28 Sep 2026	Claude Sonnet 5.5	55.5	Agentic coding CursorBench 4.0
22 Sep 2026	Claude Opus 5.5	57.8	Agentic coding CursorBench 4.0
21 Sep 2026	Grok 4.7	46.3	CursorBench 4.0 (Software engineering)

3 system cards report it, from 2 labs

Besides launch posts, we read the system cards, model cards and technical reports labs publish with a release. They report far more, including safety and frontier-risk tests.

Date	Document	Score	As the document writes it	Section
28 Sep 2026	Claude Sonnet 5.5 system card	55.5	CursorBench 4.0 at max effort, measured by Cursor (8.8, Figure 8.8.A)	Capabilities
22 Sep 2026	Claude Opus 5.5 system card	57.8	CursorBench 4.0 at max effort, measured by Cursor (56.0% at xhigh and high, 52.5% at medium) (8.8, Figure 8.8.A)	Capabilities
21 Sep 2026	Grok 4.7 model card	46.3	CursorBench 4.0, xhigh, value from text (2.1). High = 43.9%. Charts have no printed values	Capabilities

Versions

We switch the top row to each new version when it comes out. Open it for one row per version. An old version stays in play while people still test it, so whoever first takes it past 85% gets credit for closing it out.

Cursor xAI Anthropic

CursorBench 4.0 5 leaders