Best LLMs for coding: open and proprietary

Open and proprietary models together on one 0–100 scale (coding). Switch to local models to see only what you can download and run.

40 benchmarks4 sourcesUpdated 3 Oct 2026Reference scale 2026-Q4How the score is built

Local models Overall

General Coding Agents Math Science Reasoning

Models ranked by coding llmrun Score
#	Model	llmrun Score	Coding	Agents	Math	Science	Reasoning	Benchmarks	VRAM
1	GPT 6 Astra Max19 benchmarks		83	72	96	57	86	19	–
2	Claude Fable 5.1 Max15 benchmarks		79	–	95	57	75	15	–
3	GPT 6 Sol Max18 benchmarks		79	–	94	56	75	18	–
4	Claude Fable 5 Max20 benchmarks		78	62	94	54	74	20	–
5	Claude Fable 5.19 benchmarks		78	–	–	56	74	9	–
6	Claude Opus 59 benchmarks		78	–	–	54	74	9	–
7	Claude Opus 5.5 Max17 benchmarks		78	–	96	58	82	17	–
8	GPT 5.6 Sol Max21 benchmarks		78	70	91	57	77	21	–
9	Claude Opus 5 Max22 benchmarks		78	72	91	55	75	22	–
10	Claude Sonnet 5.53 benchmarks	—	75	–	–	57	–	3	–
11	GPT 5.6 Sol9 benchmarks		75	–	–	55	72	9	–
12	Claude Sonnet 5.5 Max13 benchmarks		75	–	93	58	81	13	–
13	GPT 5.59 benchmarks		74	–	–	53	72	9	–
14	GPT 6.1 Sol3 benchmarks	—	73	–	–	54	–	3	–
15	GPT 5.6 Terra Max20 benchmarks		73	67	87	55	71	20	–
16	Kimi K3open1674 GB · 26 benchmarks		73	61	81	52	63	26	1674 GB
17	GPT 5.6 Terra10 benchmarks		70	–	–	53	65	10	–
18	GPT 5.3 Codex2 benchmarks	—	70	–	–	–	–	2	–
19	GPT 5.6 Luna Max20 benchmarks		69	63	82	48	62	20	–
20	GPT 6 Luna Max18 benchmarks		69	–	82	47	60	18	–
21	GPT 5.4 (Mar 05, 2026)20 benchmarks		69	57	78	51	67	20	–
22	Gemini 3.8 Flash6 benchmarks		69	–	–	45	65	6	–
23	Claude Opus 4.77 benchmarks		68	75	–	–	54	7	–
24	Claude Opus 4.7 Max15 benchmarks		67	–	73	42	64	15	–
25	GPT 6 Luna5 benchmarks		67	–	–	47	55	5	–
26	GLM 5.3open456 GB · 21 benchmarks		67	68	72	49	58	21	456 GB
27	Muse Spark 1.19 benchmarks		66	59	69	48	–	9	–
28	Grok 4.516 benchmarks		65	–	66	47	60	16	–
29	Claude Opus 4.68 benchmarks		64	64	–	–	–	8	–
30	Claude Opus 4.8 Max19 benchmarks		64	55	82	48	66	19	–
31	GPT 6.1 Sol Max13 benchmarks		64	–	96	56	83	13	–
32	DeepSeek V4 Pro 0813open991 GB · 21 benchmarks		63	–	71	46	65	21	991 GB
33	GPT 5.2 Codex8 benchmarks		62	48	–	–	–	8	–
34	Gemini 3 Pro Preview17 benchmarks		62	55	60	46	54	17	–
35	Qwen3.7 Max17 benchmarks		59	–	68	43	58	17	–
36	DeepSeek V4 Flash 0731open183 GB · 21 benchmarks		59	–	71	45	62	21	183 GB
37	GPT 5.1 (Nov 13, 2025)14 benchmarks		59	–	–	–	50	14	–
38	Muse Spark 1.27 benchmarks		59	–	–	49	62	7	–
39	GPT 5 (Aug 07, 2025)20 benchmarks		58	–	62	40	48	20	–
40	GLM 5.2open456 GB · 24 benchmarks		58	60	68	47	52	24	456 GB
41	Kimi K2.6open620 GB · 18 benchmarks		57	–	63	43	51	18	620 GB
42	Gemini 3.7 Flash15 benchmarks		56	–	74	48	70	15	–
43	GPT 5.4 2026 03.056 benchmarks		56	–	–	–	33	6	–
44	O3 Pro (Jun 10, 2025)6 benchmarks		55	–	–	–	44	6	–
45	Grok 4.2011 benchmarks		55	49	–	–	63	11	–
46	DeepSeek V4.1 Flashopen458 GB · 10 benchmarks		54	–	76	45	57	10	458 GB
47	MiMo V2.5 Proopen615 GB · 6 benchmarks		54	–	–	38	42	6	615 GB
48	Kimi K2.7 Codeopen620 GB · 19 benchmarks		53	46	61	40	50	19	620 GB
49	GLM 5.1open457 GB · 13 benchmarks		53	54	57	40	46	13	457 GB
50	O3 (Apr 16, 2025)19 benchmarks		52	–	52	35	46	19	–
51	Gemini 3.6 Flash16 benchmarks		52	–	66	45	63	16	–
52	GLM 5.3 Flashopen195 GB · 17 benchmarks		52	63	62	44	37	17	195 GB
53	MiniMax M3open257 GB · 19 benchmarks		52	40	46	41	38	19	257 GB
54	Qwen3.8 27Bopen17.4 GB · 9 benchmarks		51	–	57	37	50	9	17.4 GB
55	GPT 5 Mini (Aug 07, 2025)15 benchmarks		50	–	57	–	42	15	–
56	Grok Build 0.15 benchmarks		50	–	–	41	–	5	–
57	DeepSeek V4 Proopen960 GB · 22 benchmarks		49	–	57	44	49	22	960 GB
58	Grok 4 (Jul 09)12 benchmarks		49	30	–	–	48	12	–
59	Kimi K2.5open620 GB · 20 benchmarks		49	37	–	40	45	20	620 GB
60	GLM 5open457 GB · 15 benchmarks		49	45	–	–	40	15	457 GB
61	O4 Mini (Apr 16, 2025)19 benchmarks		49	–	52	34	42	19	–
62	MiniMax M2.5open138 GB · 11 benchmarks		47	37	–	–	45	11	138 GB
63	Qwen3.6 Plus15 benchmarks		47	–	55	38	41	15	–
64	DeepSeek V3.1 Terminusopen415 GB · 5 benchmarks		47	–	–	31	39	5	415 GB
65	Gemma 4 31B ITopen20.4 GB · 11 benchmarks		47	–	–	32	40	11	20.4 GB
66	DeepSeek R1 0528open415 GB · 11 benchmarks		46	–	44	–	32	11	415 GB
67	DeepSeek V3.2 Expopen415 GB · 10 benchmarks		46	–	–	30	44	10	415 GB
68	Step 3.7 Flashopen133 GB · 3 benchmarks	—	46	–	–	30	–	3	133 GB
69	DeepSeek V4 Flashopen175 GB · 10 benchmarks		46	–	–	36	43	10	175 GB
70	Mercury 24 benchmarks		45	–	–	29	–	4	–
71	Qwen3.6 35B A3Bopen22.0 GB · 12 benchmarks		44	–	49	35	38	12	22.0 GB
72	Inklingopen572 GB · 21 benchmarks		43	–	52	40	53	21	572 GB
73	Gemma 4 26B A4B ITopen16.1 GB · 9 benchmarks		43	–	–	30	33	9	16.1 GB
74	Claude Sonnet 4 (May 14, 2025)4 benchmarks		43	–	–	–	–	4	–
75	Gemini 3.5 Flash Lite11 benchmarks		43	–	46	–	43	11	–
76	Kimi K2 Thinkingopen620 GB · 9 benchmarks		43	34	–	–	–	9	620 GB
77	MiMo V2.5open187 GB · 4 benchmarks		43	–	–	33	–	4	187 GB
78	NVIDIA Nemotron 3 Ultra 550B A55B BF16open370 GB · 14 benchmarks		43	–	52	37	46	14	370 GB
79	Qwen3.7 Plus9 benchmarks		42	–	–	40	46	9	–
80	O3 Mini (Jan 31, 2025)18 benchmarks		42	–	47	32	31	18	–
81	Qwen3 235B A22B Thinking 2507open142 GB · 13 benchmarks		42	–	–	34	37	13	142 GB
82	Qwen3.6 27Bopen17.4 GB · 12 benchmarks		42	–	53	36	37	12	17.4 GB
83	GPT 5 Nano (Aug 07, 2025)15 benchmarks		42	–	48	–	25	15	–
84	Mistral Medium 26049 benchmarks		41	–	–	30	33	9	–
85	MiniMax M2.7open138 GB · 6 benchmarks		41	–	–	35	–	6	138 GB
86	Qwen3 235B A22B Instruct 2507open142 GB · 8 benchmarks		41	–	–	–	31	8	142 GB
87	Kimi K2 Instructopen620 GB · 10 benchmarks		41	28	–	–	25	10	620 GB
88	GLM 4.7open216 GB · 13 benchmarks		41	32	51	36	35	13	216 GB
89	Qwen3 235B A22Bopen142 GB · 10 benchmarks		41	–	–	–	31	10	142 GB
90	GPT OSS 20Bopen12.9 GB · 11 benchmarks		40	–	–	24	26	11	12.9 GB
91	Ring 2.6 1Topen677 GB · 3 benchmarks	—	40	–	–	33	–	3	677 GB
92	Qwen3 Coder 480B A35B Instructopen289 GB · 6 benchmarks	—	40	31	–	–	–	6	289 GB
93	MiMo v2 Flashopen186 GB · 3 benchmarks	—	40	–	–	17	–	3	186 GB
94	Kimi K2 Instruct 0905open620 GB · 6 benchmarks		40	–	–	–	–	6	620 GB
95	DeepSeek R1open415 GB · 14 benchmarks		39	–	40	28	29	14	415 GB
96	DeepSeek v3 0324open415 GB · 13 benchmarks		39	–	33	27	24	13	415 GB
97	GPT 4.1 (Apr 14, 2025)17 benchmarks		39	–	34	26	26	17	–
98	Claude 3.5 Sonnet (Oct 22, 2024)11 benchmarks		38	26	23	21	29	11	–
99	GLM 4.5open216 GB · 7 benchmarks		38	32	–	–	–	7	216 GB
100	Grok 3 Mini Beta5 benchmarks		38	–	48	–	–	5	–
101	Grok Code Fast 13 benchmarks	—	38	–	–	–	–	3	–
102	Qwen3.5 27Bopen17.4 GB · 6 benchmarks		38	–	–	–	41	6	17.4 GB
103	GPT OSS 120Bopen70.5 GB · 20 benchmarks		37	14	–	31	30	20	70.5 GB
104	DeepSeek v3open415 GB · 10 benchmarks		37	–	26	22	21	10	415 GB
105	GPT 4.5 Preview (Feb 27, 2025)11 benchmarks		36	–	33	27	28	11	–
106	GLM 4.6open215 GB · 6 benchmarks		36	31	–	29	–	6	215 GB
107	NVIDIA Nemotron 3 Super 120B A12B BF16open74.7 GB · 4 benchmarks		36	–	–	27	–	4	74.7 GB
108	Gemini 2.5 Flash Lite Preview Thinking (Jun 17)5 benchmarks		35	–	–	–	24	5	–
109	DeepSeek V3.2open415 GB · 13 benchmarks		35	36	–	–	43	13	415 GB
110	Qwen3 Coder Nextopen48.2 GB · 3 benchmarks	—	35	–	–	23	–	3	48.2 GB
111	Qwen3 32Bopen20.3 GB · 9 benchmarks		33	–	–	26	26	9	20.3 GB
112	Mistral Large 3 675B Instruct 2512open446 GB · 5 benchmarks		33	–	–	26	25	5	446 GB
113	Mistral Medium 25086 benchmarks		33	–	–	24	26	6	–
114	GPT 4.1 Mini (Apr 14, 2025)16 benchmarks		32	–	36	27	26	16	–
115	Trinity Large Thinkingopen263 GB · 3 benchmarks	—	32	–	–	26	–	3	263 GB
116	Claude 3.5 Haiku (Oct 22, 2024)9 benchmarks		29	–	20	13	–	9	–
117	GPT 4o (Nov 20, 2024)10 benchmarks		25	–	21	17	24	10	–
118	Llama 4 Maverick 17B 128E Instructopen265 GB · 17 benchmarks		25	–	29	26	23	17	265 GB
119	Llama 3.3 70B Instructopen46.6 GB · 13 benchmarks		22	–	19	17	21	13	46.6 GB
120	GPT 4.1 Nano (Apr 14, 2025)13 benchmarks		21	–	30	17	17	13	–

The range after each score is a 90% interval: given the boards a model has results on, its true score is very likely inside it. Short ranges mean many agreeing results; long ranges mean few or conflicting ones.

Why some models are missing. A model gets a score only with results on at least 4 benchmarks across at least 2 categories, one of them reasoning or coding. Skill scores need at least 2 benchmarks in that skill; otherwise they show as –.

VRAM is llmrun's estimate at Q4_K_M (or the smallest quantization we track) for the weights plus a working context. Proprietary models can't run locally, so the VRAM filter hides them.

How the llmrun Score is built · llmrun does not run these benchmarks; scores are aggregated from public sources.