benchgap
Coding

cursorBench41 leaderboard

As of 2026-11-21, the highest measured score on cursorBench41 is 65.4% by Gemini 4 Neon. 80 more models have estimated scores, calibrated from the benchmarks they were measured on.

Measured scores: benchlm.ai.

#ModelScoreSource
1Gemini 4 Neon65.4%measured
2Gemini 4 Carbon61.5%measured
3Gemini 4 Hydrogen61.4%measured
4Claude Fable 5.261.3%measured
5GPT-6.2 Astra59.8%estimated ± 6.0 pp, medium confidence
6GPT-6.2 Sol59.6%estimated ± 6.0 pp, medium confidence
7Muse Wonder Full59.0%estimated ± 3.7 pp, high confidence
8Claude Opus 5.558.1%measured
9Gemini 4 Argon56.5%measured
10Claude Sonnet 5.554.4%measured
11Claude Fable 553.7%estimated ± 3.7 pp, high confidence
12MiMo-V2.6-Pro52.9%estimated ± 3.7 pp, high confidence
13Claude Fable 5.152.8%estimated ± 2.0 pp, high confidence
14Kimi K3.1 Pro52.2%estimated ± 3.7 pp, high confidence
15Muse Wonder51.3%estimated ± 2.0 pp, high confidence
16Hy451.3%estimated ± 6.0 pp, medium confidence
17Grok 4.748.6%measured
18Claude Haiku 5.548.4%measured
19DeepSeek V4.1 Flash 101748.2%estimated ± 2.0 pp, medium confidence
20Claude Opus 547.5%estimated ± 2.0 pp, high confidence
21Kimi K345.8%estimated ± 3.7 pp, high confidence
22Step 5 Preview45.1%estimated ± 3.7 pp, high confidence
23Muse Spark 1.145.1%estimated ± 3.7 pp, high confidence
24SWE-245.1%estimated ± 6.0 pp, low confidence
25Gemini 3.1 Pro45.0%estimated ± 3.7 pp, high confidence
26SWE-2 Flash45.0%estimated ± 6.0 pp, low confidence
27Muse Spark Open44.9%estimated ± 3.7 pp, high confidence
28GPT-6 Sol44.8%estimated ± 3.7 pp, high confidence
29Muse Spark 1.244.8%estimated ± 3.7 pp, high confidence
30Gemini 3.7 Flash44.8%estimated ± 3.7 pp, high confidence
31Gemini 3.8 Flash44.8%estimated ± 3.7 pp, high confidence
32GPT-6 Astra44.8%estimated ± 3.7 pp, high confidence
33GPT-5.544.8%estimated ± 3.7 pp, high confidence
34GPT-6.1 Terra44.8%estimated ± 3.7 pp, high confidence
35Grok 4.544.8%estimated ± 3.7 pp, high confidence
36GPT-6 Luna44.8%estimated ± 3.7 pp, medium confidence
37Claude Opus 4.844.8%estimated ± 3.7 pp, medium confidence
38GPT-6.1 Sol44.8%estimated ± 3.7 pp, medium confidence
39Ling 3.1 Flash44.8%estimated ± 3.7 pp, medium confidence
40Gemini 3.5 Flash44.8%estimated ± 3.7 pp, medium confidence
41Gemini 3.6 Flash44.8%estimated ± 3.7 pp, medium confidence
42MiniMax M3.1 Flash44.8%estimated ± 3.7 pp, medium confidence
43Inkling 2 Small44.8%estimated ± 3.7 pp, medium confidence
44GPT-5.4 mini44.8%estimated ± 3.7 pp, medium confidence
45Qwen3.8 Max Preview44.8%estimated ± 3.7 pp, medium confidence
46Kimi K2.644.8%estimated ± 3.7 pp, medium confidence
47MiMo-V2.6-Flash44.8%estimated ± 3.7 pp, medium confidence
48GLM-5.244.8%estimated ± 3.7 pp, medium confidence
49DeepSeek V4 Pro44.8%estimated ± 3.7 pp, medium confidence
50MiMo-V2.5-Pro44.8%estimated ± 3.7 pp, medium confidence
51Qwen3.8-Flash-Next44.8%estimated ± 3.7 pp, medium confidence
52DeepSeek V4 Flash 073144.8%estimated ± 3.7 pp, medium confidence
53MiniMax M2.744.8%estimated ± 3.7 pp, medium confidence
54Inkling-Small44.8%estimated ± 3.7 pp, medium confidence
55Qwen3.7 Max44.8%estimated ± 3.7 pp, medium confidence
56Hy344.8%estimated ± 3.7 pp, medium confidence
57Hy3 Preview44.8%estimated ± 3.7 pp, medium confidence
58Beam 1.0 501B-A23B44.8%estimated ± 3.7 pp, medium confidence
59Grok 4.344.8%estimated ± 3.7 pp, medium confidence
60Quasar 438B44.8%estimated ± 3.7 pp, medium confidence
61Solar Pro 444.8%estimated ± 3.7 pp, medium confidence
62Kimi K2.7 Code44.8%estimated ± 3.7 pp, medium confidence
63Nemotron 3.5 Super VL 120B A12B44.8%estimated ± 3.7 pp, medium confidence
64Nemotron 3.5 Super VL 120B A12B NVFP444.8%estimated ± 3.7 pp, medium confidence
65GPT-5.4 nano44.8%estimated ± 3.7 pp, medium confidence
66MiniMax M344.8%estimated ± 3.7 pp, medium confidence
67Inkling44.8%estimated ± 3.7 pp, medium confidence
68Qwen3.8-27B44.8%estimated ± 3.7 pp, medium confidence
69Gemini 2.5 Pro44.8%estimated ± 3.7 pp, medium confidence
70Qwen3.7 Plus44.8%estimated ± 3.7 pp, medium confidence
71Apodex 1.144.8%estimated ± 3.7 pp, medium confidence
72Apodex 1.1 Mini44.8%estimated ± 3.7 pp, medium confidence
73Gemma 4 31B44.8%estimated ± 3.7 pp, medium confidence
74Muse Glimmer 30B44.8%estimated ± 3.7 pp, medium confidence
75GLM-5.144.8%estimated ± 3.7 pp, medium confidence
76Qwen3.5 397B44.8%estimated ± 3.7 pp, medium confidence
77Ling 3.0 Flash VL44.8%estimated ± 3.7 pp, medium confidence
78Step 3.7 Flash44.8%estimated ± 3.7 pp, medium confidence
79Qwen3.6-27B44.8%estimated ± 3.7 pp, medium confidence
80Grok 4.644.3%measured
81GLM-5.344.0%estimated ± 2.0 pp, high confidence
82GPT-5.6 Terra42.5%estimated ± 2.0 pp, medium confidence
83Muse Spark 1.341.9%measured
84GPT-5.6 Sol40.9%measured
85Mistral Large 440.2%estimated ± 6.0 pp, low confidence
86Qwen3.8 Max38.4%estimated ± 4.3 pp, medium confidence
87GLM-5.3-Flash37.9%estimated ± 2.6 pp, medium confidence
88GPT-5.6 Luna37.3%estimated ± 2.0 pp, medium confidence
89Claude Sonnet 535.7%estimated ± 2.0 pp, medium confidence
90SWE-1.734.6%estimated ± 6.0 pp, low confidence
91Inkling 2 Mini34.0%estimated ± 6.0 pp, low confidence
92Composer 2.528.5%estimated ± 2.6 pp, medium confidence

All leaderboards

Agentic · terminal

Agentic · tools

Coding

Math

Knowledge & reasoning

Instruction following

Multilingual

Vision & documents

Long context

General