BenchLeader
Menu
Best LLMsChoose a modelCoding plansLeaderboardsModelsBenchmarksProvidersReleases7

CursorBench leaderboard

As of 6 Nov 2026, the top score on CursorBench is 57.8% by Claude Opus 5.5 (Anthropic), ahead of Claude Sonnet 5.5 at 55.5%, across 82 model configurations with a published result (Cursor, result dated 22 Sep 2026). Coding tasks run in Cursor's agent with the reasoning level and cost per task recorded. CursorBench is shown for reference only and does not count toward the BenchLeader Index.

Key facts (as of 6 Nov 2026)

  • Claude Opus 5.5 holds the top CursorBench score at 57.8% (Cursor, 22 Sep 2026).
  • Claude Sonnet 5.5 is second on CursorBench at 55.5% (Cursor, 28 Sep 2026).
  • CursorBench: Cursor's internal coding-agent evaluation, with reasoning level and cost (Cursor).
  • 82 model configurations have a published CursorBench result on BenchLeader (BenchLeader, 6 Nov 2026).
  • CursorBench does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).

Cursor's internal coding-agent evaluation, with reasoning level and cost.

Published by
Cursordata via Epoch AI Benchmarking Hub
Category
Coding
Index weight
Reference only
Models
82
Data as of
6 Nov 2026

CC BY 4.0 — Epoch AI, ‘AI Benchmarking Hub’, epoch.ai/benchmarks. Mirrored boards credit their original publishers.

What the test looks like

Coding tasks run in Cursor's agent with the reasoning level and cost per task recorded.

How it is scored

Score, as published by Cursor.

What to keep in mind

Published by a coding-tool vendor; harness is theirs.

Top 15
  1. 1Claude Opus 5.5 (max)57.8%
  2. 2Claude Opus 5.5 (xhigh)56.0%
  3. 3Claude Opus 5.5 (high)56.0%
  4. 4Claude Sonnet 5.5 (max)55.5%
  5. 5Claude Sonnet 5.5 (xhigh)53.1%
  6. 6Gemini 4 Argon (high)52.9%
  7. 7Gemini 4 Argon (medium)52.8%
  8. 8Claude Opus 5.5 (medium)52.5%
  9. 9Claude Fable 5.1 (max)51.8%
  10. 10Claude Fable 5.1 (xhigh)51.6%
  11. 11Gemini 4 Argon (low)50.6%
  12. 12Muse Wonder (max)50.4%
  13. 13Claude Fable 5.1 (high)49.2%
  14. 14Gemini 4 Argon (minimal)48.8%
  15. 15Claude Haiku 5.5 (max)48.4%
82 of 82
#
1Claude Opus 5.5max57.8%69.32026-09-22
2Claude Opus 5.5xhigh56.0%66.32026-09-22
3Claude Opus 5.5high56.0%65.82026-09-22
4Claude Sonnet 5.5max55.5%65.42026-09-28
5Claude Sonnet 5.5xhigh53.1%63.42026-09-28
6Gemini 4 Argonhigh52.9%66.02026-09-30
7Gemini 4 Argonmedium52.8%–2026-09-30
8Claude Opus 5.5medium52.5%62.02026-09-22
9Claude Fable 5.1max51.8%67.22026-09-01
10Claude Fable 5.1xhigh51.6%65.42026-09-01
11Gemini 4 Argonlow50.6%–2026-09-30
12Muse Wondermax50.4%62.32026-10-12
13Claude Fable 5.1high49.2%64.02026-09-01
14Gemini 4 Argonminimal48.8%–2026-09-30
15Claude Haiku 5.5max48.4%57.02026-10-07
16Muse Wonderxhigh48.2%–2026-10-12
17Claude Sonnet 5.5high47.8%58.62026-09-28
18DeepSeek V4.1 Flash 1017max47.7%59.22026-10-17
19Claude Fable 5.1medium46.8%62.32026-09-01
20Claude Opus 5max46.6%65.42026-07-24
21Grok 4.7xhigh46.3%58.32026-09-21
22DeepSeek V4.1 Flash 1017medium46.3%–2026-10-17
23Claude Opus 5xhigh46.1%60.22026-07-24
24Muse Wonderhigh46.1%56.92026-10-12
25DeepSeek V4.1 Flash 1017high45.6%51.72026-10-17
26Muse Wondermedium45.2%–2026-10-12
27Claude Fable 5.1low45.1%59.32026-09-01
28DeepSeek V4.1 Flash 1017low44.9%–2026-10-17
29Claude Opus 5high44.7%64.32026-07-24
30Claude Haiku 5.5xhigh44.3%57.32026-10-07
31Grok 4.7high43.9%57.72026-09-21
32Claude Opus 5.5low43.7%59.52026-09-22
33Claude Opus 5medium43.3%57.42026-07-24
34GLM 5.3max42.6%59.02026-08-14
35Claude Haiku 5.5high42.3%55.82026-10-07
36GPT-5.6 Solmax41.7%65.42026-07-09
37Muse Wonderlow41.7%–2026-10-12
38Muse Spark 1.3max41.6%62.02026-09-02
39Grok 4.7medium41.6%–2026-09-21
40Grok 4.6xhigh41.4%57.92026-08-12
41GPT-5.6 Terramax41.3%60.72026-07-09
42Claude Opus 5low40.7%57.02026-07-24
43Grok 4.6high40.4%60.52026-08-12
44Gemini 3.8 Flashhigh39.6%62.02026-09-02
45Claude Sonnet 5.5medium39.2%51.22026-09-28
46GLM 5.3high38.0%–2026-08-14
47GPT-5.6 Solxhigh37.7%64.22026-07-09
48Muse Spark 1.3xhigh37.5%58.52026-09-02
49Gemini 3.8 Flashmedium37.3%60.72026-09-02
50Claude Haiku 5.5medium36.9%50.52026-10-07
51GLM 5.3 Flashmax36.8%54.22026-08-20
52Grok 4.6medium36.1%59.62026-08-12
53GPT-5.6 Lunamax35.9%56.32026-07-09
54Claude Sonnet 5.5low35.8%51.32026-09-28
55GPT-5.6 Solhigh35.7%63.62026-07-09
56Muse Wonderminimal34.2%–2026-10-12
57Claude Sonnet 5max34.1%53.82026-06-30
58GPT-5.6 Terraxhigh33.6%60.32026-07-09
59Grok 4.6low33.4%55.22026-08-12
60Muse Spark 1.3high33.4%–2026-09-02

More coding benchmarks · All LLM benchmarks

Related benchmarks

Top head-to-heads on CursorBench

Cite as: BenchLeader, “CursorBench leaderboard”, data as of 6 Nov 2026, https://www.benchleader.com/benchmarks/cursorbench.

CursorBench: questions

What does CursorBench measure?
Coding tasks run in Cursor's agent with the reasoning level and cost per task recorded. Scores are reported in percent of tasks solved; higher is better.
What is the top score on CursorBench?
As of 6 Nov 2026, the highest score on CursorBench is 57.8% by Claude Opus 5.5 (result dated 22 Sep 2026), ahead of Claude Sonnet 5.5 at 55.5%. Scores come from Cursor and are re-read daily.
Which AI model leads CursorBench?
Claude Opus 5.5 leads CursorBench with 57.8% as of 6 Nov 2026 (result dated 22 Sep 2026), ahead of Claude Sonnet 5.5 at 55.5%.
How many models have CursorBench results?
82 model configurations have a CursorBench result on BenchLeader, all taken from Cursor via Epoch AI Benchmarking Hub.
Who runs CursorBench and how often is it updated?
CursorBench is published by Cursor. BenchLeader re-reads the published results once a day and records the date each result was published.
Does CursorBench count toward the BenchLeader Index?
No. CursorBench is shown for reference but left out of the composite index.