Best LLMs: open and proprietary

Open and proprietary models together on one 0–100 scale (overall capability). 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 overall llmrun Score
#	Model	Coding	Agents	Math	Science	Reasoning	Benchmarks	VRAM
1	GPT 6 Astra6 benchmarks	–	87	–	–	74	6	–
2	GPT 6 Astra Max19 benchmarks	83	72	96	57	86	19	–
3	Claude Opus 5.5 Max17 benchmarks	78	–	96	58	82	17	–
4	GPT 6.1 Sol Max13 benchmarks	64	–	96	56	83	13	–
5	Claude Fable 5.1 Max15 benchmarks	79	–	95	57	75	15	–
6	Claude Sonnet 5.5 Max13 benchmarks	75	–	93	58	81	13	–
7	GPT 6 Sol Max18 benchmarks	79	–	94	56	75	18	–
8	GPT 5.6 Sol Max21 benchmarks	78	70	91	57	77	21	–
9	Claude Fable 5.19 benchmarks	78	–	–	56	74	9	–
10	GPT 5.6 Sol Promax6 benchmarks	–	–	–	–	79	6	–
11	Claude Opus 5 Max22 benchmarks	78	72	91	55	75	22	–
12	Claude Opus 59 benchmarks	78	–	–	54	74	9	–
13	Claude Opus 5.54 benchmarks	–	–	–	59	74	4	–
14	Claude Fable 5 Max20 benchmarks	78	62	94	54	74	20	–
15	Claude Fable 56 benchmarks	–	–	–	–	73	6	–
16	GPT 5.5 Pro7 benchmarks	–	–	90	–	70	7	–
17	GPT 5.6 Sol9 benchmarks	75	–	–	55	72	9	–
18	GPT 5.59 benchmarks	74	–	–	53	72	9	–
19	Claude Opus 4.77 benchmarks	68	75	–	–	54	7	–
20	GPT 5.6 Terra Max20 benchmarks	73	67	87	55	71	20	–
21	Gemini 3.1 Pro Preview8 benchmarks	–	–	–	–	68	8	–
22	GPT 5.4 Pro (Mar 05, 2026)8 benchmarks	–	–	83	56	74	8	–
23	Gemini 3.8 Flash6 benchmarks	69	–	–	45	65	6	–
24	GPT 6 Sol4 benchmarks	–	–	–	54	68	4	–
25	Claude Opus 4.85 benchmarks	–	64	–	–	56	5	–
26	GPT 5.6 Terra10 benchmarks	70	–	–	53	65	10	–
27	Grok 4.65 benchmarks	–	72	–	–	–	5	–
28	Claude Opus 4.8 Max19 benchmarks	64	55	82	48	66	19	–
29	Kimi K3open1674 GB · 26 benchmarks	73	61	81	52	63	26	1674 GB
30	Claude Opus 4.68 benchmarks	64	64	–	–	–	8	–
31	Claude Sonnet 55 benchmarks	–	62	–	–	–	5	–
32	Gemini 3.7 Flash15 benchmarks	56	–	74	48	70	15	–
33	GPT 5.4 (Mar 05, 2026)20 benchmarks	69	57	78	51	67	20	–
34	GPT 5.6 Luna Max20 benchmarks	69	63	82	48	62	20	–
35	Muse Spark 1.3 Max6 benchmarks	–	–	78	54	61	6	–
36	Qwen3.8 Max10 benchmarks	–	–	78	–	62	10	–
37	GPT 6 Luna Max18 benchmarks	69	–	82	47	60	18	–
38	Muse Spark 1.27 benchmarks	59	–	–	49	62	7	–
39	Claude Sonnet 5 Max13 benchmarks	–	–	77	38	57	13	–
40	Claude Opus 4.7 Max15 benchmarks	67	–	73	42	64	15	–
41	GLM 5.3open456 GB · 21 benchmarks	67	68	72	49	58	21	456 GB
42	Muse Spark 1.36 benchmarks	–	–	76	–	53	6	–
43	Gemini 3.5 Flash4 benchmarks	–	–	–	–	65	4	–
44	DeepSeek V4 Pro 0813open991 GB · 21 benchmarks	63	–	71	46	65	21	991 GB
45	Qwen3.8 Max (Sep 02)6 benchmarks	–	–	72	–	–	6	–
46	Muse Spark 1.19 benchmarks	66	59	69	48	–	9	–
47	Gemini 3 Flash Preview7 benchmarks	–	–	–	–	57	7	–
48	DeepSeek V4 Flash 0731open183 GB · 21 benchmarks	59	–	71	45	62	21	183 GB
49	Claude Opus 4.6 Max12 benchmarks	–	–	71	45	55	12	–
50	DeepSeek V4.1 Flashopen458 GB · 10 benchmarks	54	–	76	45	57	10	458 GB
51	Gemini 3.6 Flash16 benchmarks	52	–	66	45	63	16	–
52	Grok 4.2011 benchmarks	55	49	–	–	63	11	–
53	Grok 4.516 benchmarks	65	–	66	47	60	16	–
54	GPT 5.6 Luna8 benchmarks	–	–	–	48	58	8	–
55	GPT 5.2 (Dec 11, 2025)12 benchmarks	–	–	65	–	61	12	–
56	Qwen3.7 Max17 benchmarks	59	–	68	43	58	17	–
57	Claude Sonnet 4.6 Max11 benchmarks	–	–	70	34	59	11	–
58	Claude Opus 4.5 (Nov 01, 2025)9 benchmarks	–	52	–	–	55	9	–
59	Grok 4.79 benchmarks	–	–	63	49	63	9	–
60	GPT 6 Luna5 benchmarks	67	–	–	47	55	5	–
61	Claude Sonnet 4.68 benchmarks	–	43	–	–	60	8	–
62	Gemini 3 Pro Preview17 benchmarks	62	55	60	46	54	17	–
63	GLM 5.2open456 GB · 24 benchmarks	58	60	68	47	52	24	456 GB
64	Qwen3.6 Max Preview8 benchmarks	–	–	–	–	52	8	–
65	GPT 5.2 Codex8 benchmarks	62	48	–	–	–	8	–
66	Kimi K2.6open620 GB · 18 benchmarks	57	–	63	43	51	18	620 GB
67	GPT 5 Pro (Oct 06, 2025)5 benchmarks	–	–	64	–	51	5	–
68	GLM 5.3 Flashopen195 GB · 17 benchmarks	52	63	62	44	37	17	195 GB
69	Muse Spark6 benchmarks	–	–	58	47	–	6	–
70	Inkling Smallopen160 GB · 12 benchmarks	–	–	57	41	54	12	160 GB
71	GPT 5 (Aug 07, 2025)20 benchmarks	58	–	62	40	48	20	–
72	GPT 5.1 (Nov 13, 2025)14 benchmarks	59	–	–	–	50	14	–
73	Kimi K2.7 Codeopen620 GB · 19 benchmarks	53	46	61	40	50	19	620 GB
74	Grok 4.20 0309 Reasoning8 benchmarks	–	–	58	–	49	8	–
75	GLM 5.1open457 GB · 13 benchmarks	53	54	57	40	46	13	457 GB
76	Grok 4.311 benchmarks	–	–	57	41	52	11	–
77	GPT 5.4 Mini (Mar 17, 2026)12 benchmarks	–	–	60	40	47	12	–
78	DeepSeek V4 Proopen960 GB · 22 benchmarks	49	–	57	44	49	22	960 GB
79	MiMo V2.5 Proopen615 GB · 6 benchmarks	54	–	–	38	42	6	615 GB
80	Inklingopen572 GB · 21 benchmarks	43	–	52	40	53	21	572 GB
81	Qwen3.8 27Bopen17.4 GB · 9 benchmarks	51	–	57	37	50	9	17.4 GB
82	Claude Sonnet 4.5 (Sep 29, 2025)11 benchmarks	–	42	–	32	46	11	–
83	Gemini 2.5 Pro Preview (Jun 05)6 benchmarks	–	–	–	38	–	6	–
84	Qwen3.6 Plus15 benchmarks	47	–	55	38	41	15	–
85	Qwen3.7 Plus9 benchmarks	42	–	–	40	46	9	–
86	O3 (Apr 16, 2025)19 benchmarks	52	–	52	35	46	19	–
87	Grok 4 (Jul 09)12 benchmarks	49	30	–	–	48	12	–
88	O3 Pro (Jun 10, 2025)6 benchmarks	55	–	–	–	44	6	–
89	Kimi K2.5open620 GB · 20 benchmarks	49	37	–	40	45	20	620 GB
90	Qwen3.5 397B A17Bopen266 GB · 9 benchmarks	–	–	52	–	46	9	266 GB
91	GPT 5 Mini (Aug 07, 2025)15 benchmarks	50	–	57	–	42	15	–
92	MiniMax M2.5open138 GB · 11 benchmarks	47	37	–	–	45	11	138 GB
93	NVIDIA Nemotron 3 Ultra 550B A55B BF16open370 GB · 14 benchmarks	43	–	52	37	46	14	370 GB
94	DeepSeek Reasoner4 benchmarks	–	–	–	–	–	4	–
95	GLM 5open457 GB · 15 benchmarks	49	45	–	–	40	15	457 GB
96	Qwen3.6 27Bopen17.4 GB · 12 benchmarks	42	–	53	36	37	12	17.4 GB
97	Claude Opus 4 (May 14, 2025)7 benchmarks	–	–	–	26	46	7	–
98	MiniMax M3open257 GB · 19 benchmarks	52	40	46	41	38	19	257 GB
99	GPT 5.4 Nano (Mar 17, 2026)6 benchmarks	–	–	–	39	42	6	–
100	Grok 4 Fast6 benchmarks	–	–	–	–	41	6	–
101	Claude Opus 4.1 (Aug 05, 2025)6 benchmarks	–	–	–	–	46	6	–
102	Grok 4.1 Fast Reasoning6 benchmarks	–	–	–	–	50	6	–
103	Qwen3.5 Plus11 benchmarks	–	–	–	–	43	11	–
104	O4 Mini (Apr 16, 2025)19 benchmarks	49	–	52	34	42	19	–
105	DeepSeek V3.2open415 GB · 13 benchmarks	35	36	–	–	43	13	415 GB
106	Gemini 3.5 Flash Lite11 benchmarks	43	–	46	–	43	11	–
107	Qwen3.7 Flash5 benchmarks	–	–	49	–	46	5	–
108	Qwen3.5 122B A10Bopen82.6 GB · 6 benchmarks	–	–	–	26	43	6	82.6 GB
109	DeepSeek V4 Flashopen175 GB · 10 benchmarks	46	–	–	36	43	10	175 GB
110	DeepSeek V3.2 Expopen415 GB · 10 benchmarks	46	–	–	30	44	10	415 GB
111	Gemini 2.5 Pro Exp (Mar 25)6 benchmarks	–	–	–	37	–	6	–
112	Gemini 3.1 Flash Lite6 benchmarks	–	–	49	–	38	6	–
113	Kimi K2 Thinkingopen620 GB · 9 benchmarks	43	34	–	–	–	9	620 GB
114	Gemini 2.5 Pro13 benchmarks	–	29	49	37	43	13	–
115	Grok Build 0.15 benchmarks	50	–	–	41	–	5	–
116	GLM 4.7open216 GB · 13 benchmarks	41	32	51	36	35	13	216 GB
117	MiMo V2.5open187 GB · 4 benchmarks	43	–	–	33	–	4	187 GB
118	Qwen3 235B A22B Thinking 2507open142 GB · 13 benchmarks	42	–	–	34	37	13	142 GB
119	DeepSeek V3.1 Terminusopen415 GB · 5 benchmarks	47	–	–	31	39	5	415 GB
120	Qwen3.6 35B A3Bopen22.0 GB · 12 benchmarks	44	–	49	35	38	12	22.0 GB

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.