Coding

CursorBench 4.0 leaderboard

Name: benchgap: LLM benchmark scores, measured and estimated
Creator: benchgap

As of 2026-10-07, the highest measured score on CursorBench 4.0 is 57.8% by Claude Opus 5.5. 155 more models have estimated scores, calibrated from the benchmarks they were measured on.

Measured scores: benchlm.ai.

#	Model	Score	Source
1	Claude Opus 5.5	57.8%	measured
2	Claude Sonnet 5.5	55.5%	measured
3	Claude Fable 5.1	51.8%	measured
4	Claude Mythos 5	51.0%	estimated ± 3.6 pp, high confidence
5	MiMo-V2.6-Pro	50.0%	estimated ± 3.5 pp, high confidence
6	Gemini 4 Argon	49.9%	estimated ± 2.9 pp, high confidence
7	Claude Opus 5	46.6%	measured
8	Step 5 Preview	46.5%	estimated ± 3.5 pp, high confidence
9	Grok 4.7	46.3%	measured
10	Sakana Fugu-Ultra	46.0%	estimated ± 3.6 pp, high confidence
11	GPT-6 Sol	44.1%	estimated ± 3.5 pp, high confidence
12	GPT-6 Astra	42.5%	estimated ± 1.6 pp, high confidence
13	GLM-5.3-Flash	42.0%	estimated ± 2.9 pp, medium confidence
14	Qwen3.8 Max	42.0%	estimated ± 2.9 pp, medium confidence
15	Claude Fable 5	41.7%	estimated ± 1.6 pp, high confidence
16	GPT-5.6 Sol	41.7%	measured
17	Muse Spark 1.3	41.6%	measured
18	Grok 4.6	41.4%	measured
19	GPT-5.6 Terra	41.3%	measured
20	Kimi K3	41.3%	estimated ± 1.6 pp, high confidence
21	Gemini 3.7 Flash	41.1%	estimated ± 1.6 pp, high confidence
22	Hy4 preview	39.8%	estimated ± 3.6 pp, high confidence
23	Beam	39.7%	estimated ± 3.6 pp, high confidence
24	Gemini 3.8 Flash	39.6%	measured
25	Ornith-1.5-397B	39.4%	estimated ± 3.6 pp, high confidence
26	GPT-5.5	39.2%	estimated ± 1.6 pp, high confidence
27	GLM-5.3	39.0%	estimated ± 1.6 pp, high confidence
28	Claude Haiku 5.5	39.0%	estimated ± 3.5 pp, high confidence
29	Claude Opus 4.8	38.2%	estimated ± 1.6 pp, high confidence
30	GPT-6 Luna	38.1%	estimated ± 3.5 pp, high confidence
31	Qwen3.8-Omni-Flash	38.0%	estimated ± 3.6 pp, high confidence
32	GPT-6.1 Sol	37.3%	estimated ± 3.5 pp, high confidence
33	Mistral Large 4	37.3%	estimated ± 3.5 pp, high confidence
34	Claude Opus 4.7 (Adaptive)	37.3%	estimated ± 1.6 pp, high confidence
35	Ornith-1.0-397B	37.2%	estimated ± 3.6 pp, medium confidence
36	Ling 3.1 Flash	37.1%	estimated ± 3.5 pp, high confidence
37	Qwen3.8-Flash-Next	36.5%	estimated ± 1.6 pp, high confidence
38	dots3-note Preview	36.2%	estimated ± 3.6 pp, medium confidence
39	GPT-5.6 Luna	35.9%	measured
40	Grok 4.5	35.7%	estimated ± 1.6 pp, high confidence
41	Muse Spark 1.2	35.4%	estimated ± 1.6 pp, high confidence
42	Atria Dawn Preview	35.2%	estimated ± 3.6 pp, medium confidence
43	Ornith-1.5-35B-A3B	35.2%	estimated ± 3.6 pp, medium confidence
44	Claude Haiku 4.5	35.1%	estimated ± 3.6 pp, low confidence
45	Laguna S 2.1	35.0%	estimated ± 3.6 pp, medium confidence
46	Qwen3.8 Max Preview	34.8%	estimated ± 1.6 pp, high confidence
47	Sakana Fugu	34.7%	estimated ± 3.6 pp, medium confidence
48	Muse Spark 1.1	34.2%	estimated ± 1.6 pp, medium confidence
49	Claude Sonnet 5	34.1%	measured
50	GPT-5.4	33.9%	estimated ± 1.6 pp, medium confidence
51	Qwen 3.6 Max (preview)	33.4%	estimated ± 3.6 pp, medium confidence
52	Claude Opus 4.5	33.3%	estimated ± 3.6 pp, medium confidence
53	GPT-5.3 Codex	33.0%	estimated ± 3.6 pp, medium confidence
54	Gemini 3.5 Flash	32.7%	estimated ± 1.6 pp, medium confidence
55	MiMo-V2.5	32.5%	estimated ± 3.6 pp, medium confidence
56	Claude Opus 4.7	32.5%	estimated ± 4.4 pp, medium confidence
57	DeepSeek V4.1 Flash	32.4%	estimated ± 3.5 pp, medium confidence
58	GPT-5.2	32.1%	estimated ± 3.6 pp, medium confidence
59	GLM-5	31.7%	estimated ± 3.6 pp, medium confidence
60	Gemini 3.6 Flash	31.7%	estimated ± 1.6 pp, medium confidence
61	DeepSeek V4 Flash 0731	31.5%	estimated ± 1.6 pp, medium confidence
62	DeepSeek V4 Pro 0813	31.2%	estimated ± 1.6 pp, medium confidence
63	Gemini 3.1 Pro	31.2%	estimated ± 1.6 pp, medium confidence
64	MiMo-V2.6-Flash	31.2%	estimated ± 3.5 pp, medium confidence
65	GLM-5.2	31.1%	estimated ± 1.6 pp, medium confidence
66	Claude Opus 4.6	30.4%	estimated ± 3.6 pp, medium confidence
67	Qwen3.8-27B	30.4%	estimated ± 1.6 pp, medium confidence
68	MAI-Thinking-1	30.0%	estimated ± 3.6 pp, medium confidence
69	Claude Sonnet 4.6	29.5%	estimated ± 4.4 pp, low confidence
70	Grok 4.20	29.2%	estimated ± 3.6 pp, medium confidence
71	Qwen3.5 397B	28.5%	estimated ± 3.6 pp, medium confidence
72	Qwen3.7 Max	28.1%	estimated ± 1.6 pp, medium confidence
73	Ornith-1.0-35B	28.1%	estimated ± 3.6 pp, medium confidence
74	Gemini 3 Flash	28.0%	estimated ± 4.4 pp, low confidence
75	Composer 2.5	27.7%	measured
76	Laguna M.1	27.2%	estimated ± 3.6 pp, medium confidence
77	GPT-5.2-Codex	26.5%	estimated ± 4.4 pp, low confidence
78	Laguna XS 2.1	26.0%	estimated ± 3.6 pp, medium confidence
79	Ornith-1.5-9B	25.9%	estimated ± 3.6 pp, medium confidence
80	Laguna XS.2	25.0%	estimated ± 3.6 pp, medium confidence
81	Kimi K2.6	24.2%	estimated ± 1.6 pp, medium confidence
82	Quasar 438B	23.8%	estimated ± 1.6 pp, medium confidence
83	Apodex 1.1	23.4%	estimated ± 1.6 pp, medium confidence
84	Apodex 1.1 Mini	23.4%	estimated ± 1.6 pp, medium confidence
85	Kimi K2.7 Code	23.4%	estimated ± 1.6 pp, medium confidence
86	MiMo-V2.5-Pro	22.9%	estimated ± 1.6 pp, medium confidence
87	Ornith-1.0-9B	22.4%	estimated ± 3.6 pp, medium confidence
88	Qwen3.5 Flash	22.2%	estimated ± 4.4 pp, low confidence
89	Hy3	21.8%	estimated ± 1.6 pp, medium confidence
90	Hy3 Preview	21.8%	estimated ± 1.6 pp, medium confidence
91	Muse Spark	21.7%	estimated ± 1.6 pp, medium confidence
92	MiniMax M3	21.7%	estimated ± 1.6 pp, medium confidence
93	Gemini 3.1 Flash-Lite	21.4%	estimated ± 4.4 pp, low confidence
94	LongCat-Flash-Lite-Sparse	20.6%	estimated ± 3.6 pp, medium confidence
95	GPT-5.4 mini	19.8%	estimated ± 1.6 pp, medium confidence
96	GPT-5.4 nano	19.8%	estimated ± 1.6 pp, medium confidence
97	Qwen3.7 Plus	19.7%	estimated ± 1.6 pp, medium confidence
98	GLM-5.1	19.6%	estimated ± 1.6 pp, medium confidence
99	Qwen3.6 Plus	18.8%	estimated ± 1.6 pp, medium confidence
100	Qwen3.6-27B	18.3%	estimated ± 1.6 pp, medium confidence
101	Inkling-Small	17.8%	estimated ± 1.6 pp, medium confidence
102	Solar Pro 4	17.6%	estimated ± 3.5 pp, medium confidence
103	MiniMax M2.7	17.6%	estimated ± 1.6 pp, medium confidence
104	Inkling	17.2%	estimated ± 1.6 pp, medium confidence
105	Ling 3.0 Flash VL	16.9%	estimated ± 3.5 pp, medium confidence
106	Ling 3.0 Flash	16.4%	estimated ± 1.6 pp, medium confidence
107	Ling 3.0 Flash FP8	16.4%	estimated ± 1.6 pp, medium confidence
108	MiMo-V2-Flash	15.9%	estimated ± 1.6 pp, medium confidence
109	GPT-5.1	15.7%	estimated ± 1.6 pp, medium confidence
110	Gemini 3.5 Flash-Lite	15.6%	estimated ± 1.6 pp, medium confidence
111	Nemotron 3 Ultra	15.6%	estimated ± 1.6 pp, medium confidence
112	Muse Glimmer 30B	15.5%	estimated ± 1.6 pp, medium confidence
113	Mistral Medium 3.5 128B	14.3%	estimated ± 1.6 pp, medium confidence
114	Kimi K2.5	14.3%	estimated ± 1.6 pp, medium confidence
115	Kimi K2.5 (Reasoning)	14.3%	estimated ± 1.6 pp, medium confidence
116	Qwen3.5-122B-A10B	13.7%	estimated ± 1.6 pp, medium confidence
117	GLM-4.7	13.5%	estimated ± 1.6 pp, medium confidence
118	K-EXAONE 2.0	13.2%	estimated ± 3.5 pp, medium confidence
119	Gemma 4 31B	12.6%	estimated ± 1.6 pp, medium confidence
120	LLaDA2.2-flash	12.6%	estimated ± 3.6 pp, medium confidence
121	Grok 4.3	12.1%	estimated ± 1.6 pp, medium confidence
122	Qwen3.6-35B-A3B	11.9%	estimated ± 1.6 pp, medium confidence
123	A.X K2	11.7%	estimated ± 3.5 pp, medium confidence
124	o1	11.0%	estimated ± 1.6 pp, medium confidence
125	Step 3.7 Flash	10.9%	estimated ± 1.6 pp, medium confidence
126	Gemma 4 26B A4B	10.8%	estimated ± 1.6 pp, medium confidence
127	GPT-5 (high)	10.2%	estimated ± 1.6 pp, medium confidence
128	Nemotron 3 Super 100B	10.1%	estimated ± 1.6 pp, medium confidence
129	DeepSeek V3 0324	9.0%	estimated ± 3.5 pp, medium confidence
130	North Mini Code	8.8%	estimated ± 3.5 pp, medium confidence
131	o1-preview	8.7%	estimated ± 1.6 pp, medium confidence
132	Gemini 2.5 Pro	8.5%	estimated ± 1.6 pp, medium confidence
133	Mercury 2.5	8.4%	estimated ± 3.5 pp, medium confidence
134	K-Exaone	8.0%	estimated ± 1.6 pp, medium confidence
135	Gemma 4 12B	7.7%	estimated ± 1.6 pp, medium confidence
136	Granite 4.2 30B	7.6%	estimated ± 3.5 pp, medium confidence
137	GPT-OSS 120B	7.5%	estimated ± 1.6 pp, medium confidence
138	Command A+	6.6%	estimated ± 1.6 pp, medium confidence
139	Nemotron 3.5 Lightning 30B A3B NVFP4	6.3%	estimated ± 1.6 pp, medium confidence
140	Mistral Small 4	6.3%	estimated ± 1.6 pp, medium confidence
141	Mistral Small 4 (Reasoning)	6.3%	estimated ± 1.6 pp, medium confidence
142	Trinity-Large-Preview	6.0%	estimated ± 1.6 pp, medium confidence
143	Trinity-Large-Thinking	6.0%	estimated ± 1.6 pp, medium confidence
144	Ling 2.6 Flash	5.8%	estimated ± 1.6 pp, medium confidence
145	Gemini 1.5 Pro	5.4%	estimated ± 1.6 pp, medium confidence
146	DeepSeek V3	5.2%	estimated ± 1.6 pp, medium confidence
147	Granite 4.2 8B	5.0%	estimated ± 1.6 pp, medium confidence
148	GPT-4 Turbo	4.8%	estimated ± 1.6 pp, medium confidence
149	GPT-OSS 20B	4.6%	estimated ± 1.6 pp, medium confidence
150	GPT-4.1 mini	4.4%	estimated ± 1.6 pp, medium confidence
151	Mistral Large 3	4.4%	estimated ± 1.6 pp, medium confidence
152	Claude 3 Opus	4.2%	estimated ± 1.6 pp, medium confidence
153	Llama 4 Maverick	3.4%	estimated ± 1.6 pp, medium confidence
154	Celeris-1	3.0%	estimated ± 1.6 pp, medium confidence
155	Nemotron 3 Nano 30B	3.0%	estimated ± 1.6 pp, medium confidence
156	Nemotron 3 Nano Omni 30B A3B	2.8%	estimated ± 1.6 pp, medium confidence
157	Ultravox v0.6 Llama 3.3 70B	2.4%	estimated ± 1.6 pp, medium confidence
158	GPT-4o mini	2.3%	estimated ± 1.6 pp, medium confidence
159	GPT-4.1 nano	2.2%	estimated ± 1.6 pp, medium confidence
160	Gemma 3 27B	2.0%	estimated ± 1.6 pp, medium confidence
161	Gemma 4 E4B	1.8%	estimated ± 1.6 pp, medium confidence
162	Llama 4 Scout	1.6%	estimated ± 1.6 pp, medium confidence
163	LFM2.5-2.6B	1.5%	estimated ± 1.6 pp, medium confidence
164	Gemma 4 E2B	1.4%	estimated ± 1.6 pp, medium confidence
165	MiniCPM5-2B	1.0%	estimated ± 3.5 pp, medium confidence
166	Solar Pro 3	0.8%	estimated ± 3.5 pp, medium confidence
167	Granite 4.2 3B	0.8%	estimated ± 3.5 pp, medium confidence
168	Ling 3.0 Tiny	0.6%	estimated ± 3.5 pp, medium confidence

CursorBench 4.0 leaderboard

All leaderboards

Agentic · terminal

Agentic · tools

Coding

Math

Knowledge & reasoning

Instruction following

Multilingual

Vision & documents

Long context

General