Skip to content

Benchmark

CursorBench 4.0

Coding benchmarks compare models on programming tasks. Scores are percentages. Higher is better.

Models and systems measured
14
Scores recorded
69
Sources
2
IQ weight
0.81

Score distribution

41.1%
Mean
41.6%
Median
10.4%
Std dev
27.7%
10th percentile
56%
90th percentile
16%
Min
16%60%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.81 for chat models, measured over 9 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#Model or systemScore
1Claude Fable 5.2· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim60%
2Claude Opus 5.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim57.8%
3Claude Opus 5.5Cursor production agent · max effort· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.57.8%
4Claude Opus 5.5Cursor production agent · max effort · $13.43/task · 218,363 tokens · 185 steps· Reported by Cursor· via BenchBoard docs57.8%
5Claude Opus 5.5Cursor production agent · extra high effort· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.56%
6Claude Opus 5.5Cursor production agent · extra high effort · $6.98/task · 101,083 tokens · 109 steps· Reported by Cursor· via BenchBoard docs56%
7Claude Opus 5.5Cursor production agent · high effort· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.56%
8Claude Opus 5.5Cursor production agent · high effort · $3.97/task · 53,078 tokens · 68 steps· Reported by Cursor· via BenchBoard docs56%
9Claude Sonnet 5.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim55.5%
10Claude Sonnet 5.5Cursor production agent · max effort· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.55.5%
11Claude Sonnet 5.5Cursor production agent · xhigh effort· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.53.1%
12Claude Opus 5.5Cursor production agent · medium effort· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.52.5%
13Claude Opus 5.5Cursor production agent · medium effort · $2.91/task · 37,954 tokens · 54 steps· Reported by Cursor· via BenchBoard docs52.5%
14Claude Fable 5.1Cursor production agent · max effort· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.51.8%
15Claude Fable 5.1Cursor production agent · max effort · $17.28/task · 117,236 tokens · 128 steps· Reported by Cursor· via BenchBoard docs51.8%
16Claude Fable 5.1Cursor production agent · extra high effort · $13.01/task · 87,294 tokens · 101 steps· Reported by Cursor· via BenchBoard docs51.6%
17Claude Fable 5.1Cursor production agent · high effort · $9.08/task · 58,438 tokens · 77 steps· Reported by Cursor· via BenchBoard docs49.2%
18Claude Sonnet 5.5Cursor production agent · high effort· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.47.8%
19Claude Fable 5.1Cursor production agent · medium effort · $7.05/task · 45,411 tokens · 63 steps· Reported by Cursor· via BenchBoard docs46.8%
20Claude Opus 5Cursor production agent · max effort· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.46.6%
21Claude Opus 5Cursor production agent · max effort · $11.95/task · 85,384 tokens · 106 steps· Reported by Cursor· via BenchBoard docs46.6%
22Grok 4.7· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim46.3%
23Grok 4.7Cursor production agent · extra high effort· Reported by SpaceXAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.46.3%
24Grok 4.7Cursor production agent · extra high effort · $6.01/task · 70,141 tokens · 88 steps· Reported by Cursor· via BenchBoard docs46.3%
25Claude Opus 5Cursor production agent · extra high effort · $11.43/task · 80,094 tokens · 103 steps· Reported by Cursor· via BenchBoard docs46.1%
26Claude Fable 5.1Cursor production agent · low effort · $5.44/task · 34,795 tokens · 51 steps· Reported by Cursor· via BenchBoard docs45.1%
27Claude Opus 5Cursor production agent · high effort · $9/task · 61,405 tokens · 86 steps· Reported by Cursor· via BenchBoard docs44.7%
28Grok 4.7Cursor production agent · high effort· Reported by SpaceXAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.43.9%
29Grok 4.7Cursor production agent · high effort · $4.69/task · 56,382 tokens · 71 steps· Reported by Cursor· via BenchBoard docs43.9%
30Claude Opus 5.5Cursor production agent · low effort · $1.17/task · 15,811 tokens · 28 steps· Reported by Cursor· via BenchBoard docs43.7%
31Claude Opus 5Cursor production agent · medium effort · $6.94/task · 45,272 tokens · 72 steps· Reported by Cursor· via BenchBoard docs43.3%
32GPT-5.6 SolCursor production agent · max effort· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.41.7%
33GPT-5.6 SolCursor production agent · max effort · $8.23/task · 42,944 tokens · 99 steps· Reported by Cursor· via BenchBoard docs41.7%
34Grok 4.7Cursor production agent · medium effort · $3.49/task · 36,683 tokens · 60 steps· Reported by Cursor· via BenchBoard docs41.6%
35Muse Spark 1.3Cursor production agent · max effort · $2.64/task · 52,005 tokens · 98 steps· Reported by Cursor· via BenchBoard docs41.6%
36GPT-5.6 TerraCursor production agent · max effort · $5.14/task · 60,814 tokens · 107 steps· Reported by Cursor· via BenchBoard docs41.3%
37Claude Opus 5Cursor production agent · low effort · $4.87/task · 31,995 tokens · 57 steps· Reported by Cursor· via BenchBoard docs40.7%
38Gemini 3.8 FlashCursor production agent · high effort · $4.7/task · 162,565 tokens · 324 steps· Reported by Cursor· via BenchBoard docs39.6%
39Claude Sonnet 5.5Cursor production agent · medium effort· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.39.2%
40GPT-5.6 SolCursor production agent · extra high effort · $4.4/task · 24,729 tokens · 55 steps· Reported by Cursor· via BenchBoard docs37.7%
41Muse Spark 1.3Cursor production agent · extra high effort · $2.1/task · 40,891 tokens · 83 steps· Reported by Cursor· via BenchBoard docs37.5%
42Gemini 3.8 FlashCursor production agent · medium effort · $4.06/task · 128,364 tokens · 290 steps· Reported by Cursor· via BenchBoard docs37.3%
43GPT-5.6 LunaCursor production agent · max effort · $1.03/task · 87,284 tokens · 208 steps· Reported by Cursor· via BenchBoard docs35.9%
44GPT-5.6 SolCursor production agent · high effort · $2.85/task · 16,174 tokens · 41 steps· Reported by Cursor· via BenchBoard docs35.7%
45Claude Sonnet 5Cursor production agent · max effort · $7.17/task · 149,257 tokens · 140 steps· Reported by Cursor· via BenchBoard docs34.1%
46GPT-5.6 TerraCursor production agent · extra high effort · $1.81/task · 23,436 tokens · 43 steps· Reported by Cursor· via BenchBoard docs33.6%
47Muse Spark 1.3Cursor production agent · high effort · $1.66/task · 30,654 tokens · 69 steps· Reported by Cursor· via BenchBoard docs33.4%
48Grok 4.7Cursor production agent · low effort · $1.58/task · 15,677 tokens · 40 steps· Reported by Cursor· via BenchBoard docs33.1%
49GPT-5.6 LunaCursor production agent · extra high effort · $0.44/task · 40,598 tokens · 98 steps· Reported by Cursor· via BenchBoard docs33%
50Muse Spark 1.3Cursor production agent · medium effort · $1.49/task · 27,255 tokens · 64 steps· Reported by Cursor· via BenchBoard docs32.6%

Showing 1–50 of 64 results.

Not ranked

The source marks these results as descriptive, not capability scores, so we list them by name instead of ranking them.

Model or systemScore
Composer 2.5Cursor production agent · default effort · $0.68/task · 17,347 tokens · 41 steps· Reported by Cursor· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.27.7%
Grok 4.6Cursor production agent · extra high effort · $6.1/task · 49,814 tokens · 56 steps· Reported by Cursor· via BenchBoard docs41.4%
Grok 4.6Cursor production agent · high effort · $5.2/task · 41,387 tokens · 48 steps· Reported by Cursor· via BenchBoard docs40.4%
Grok 4.6Cursor production agent · low effort · $2.25/task · 16,307 tokens · 32 steps· Reported by Cursor· via BenchBoard docs33.4%
Grok 4.6Cursor production agent · medium effort · $3.48/task · 24,893 tokens · 40 steps· Reported by Cursor· via BenchBoard docs36.1%

Data credits: BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links. · Data by LLM Stats