LLM benchmarks and leaderboard

Compare benchmark results for 292 AI models across intelligence, coding, math, output speed, context windows, and API pricing. Scores are provided by Artificial Analysis and retain the reasoning configuration that was tested.

Catalogue refreshed 2026-09-04T01:12:13Z.

LLM benchmark leaderboard

Rank	Model	Lab	Open	Configuration	Intelligence	Coding	Math	Output tokens/s	Input / output per 1M tokens	Context
1	Claude Fable 5.1	Anthropic	Closed	Adaptive Reasoning, Max Effort, Default Fallback	65.7	81.6	not measured	70.3	$5 / $25	1M
2	Claude Opus 5	Anthropic	Closed	Adaptive Reasoning, Max Effort	63.1	78.0	not measured	49.0	$2.5 / $12.5	1M
3	Claude Fable 5	Anthropic	Closed	Adaptive Reasoning, Max Effort, Opus 4.8 Fallback	62.1	76.5	not measured	66.9	$3 / $18.5	1M
4	Muse Spark 1.3	Meta	Closed	max	62.1	76.3	not measured	110.8	$1.25 / $4.25	1.05M
5	GPT-5.6 Sol	OpenAI	Closed	max	60.9	77.4	not measured	80.0	$1 / $5	1.05M
6	Grok 4.6	xAI	Closed	high	60.9	76.8	not measured	59.9	$2 / $6	5K
7	Kimi K3	Moonshot AI	Open	max	59.7	76.2	not measured	37.9	$2 / $8	1.05M
8	GLM-5.3	Z.AI	Open	max	59.5	74.8	not measured	73.6	$0.4 / $1.4	1M
9	Gemini 3.8 Flash	Google	Closed	high	58.7	76.3	not measured	311.5	$0.375 / $1.88	1.05M
10	Qwen3.8 Max	Alibaba	Open	default	58.1	71.8	not measured	39.5	$1.6 / $4.8	1M
11	Qwen3.8 2.4T A95B	Alibaba	Open	default	57.7	71.9	not measured	40.2	$2 / $6	262K
12	GLM-5.3-Flash	Z.AI	Open	default	57.5	71.5	not measured	43.8	$0.015 / $0.05	1M
13	Claude Opus 4.8	Anthropic	Closed	Adaptive Reasoning, Max Effort	57.3	74.3	not measured	not measured	$0.425 / $2.12	1M
14	Muse Spark 1.2	Meta	Closed	xhigh	56.8	72.2	not measured	200.9	$1.25 / $4.25	1.05M
15	GPT-5.6 Terra	OpenAI	Closed	max	56.6	76.7	not measured	100.6	$1.5 / $2	1.05M
16	GPT-5.5	OpenAI	Closed	xhigh	56.3	74.9	not measured	not measured	$0.188 / $1.12	1.05M
17	Gemini 3.7 Flash	Google	Closed	high	56.0	76.1	not measured	291.9	$0.375 / $1.88	1.05M
18	Grok 4.5	xAI	Closed	high	55.8	72.4	not measured	50.5	$2 / $6	5K
19	Qwen3.8 Flash Next	Alibaba	Open	default	55.8	73.1	not measured	89.1	$0.12 / $0.4	262K
20	Claude Sonnet 5	Anthropic	Closed	Adaptive Reasoning, Max Effort	55.3	71.5	not measured	78.2	$1 / $5	1M
21	Claude Opus 4.7	Anthropic	Closed	Adaptive Reasoning, Max Effort	55.0	73.6	not measured	not measured	$2.5 / $12.5	1M
22	DeepSeek V4 Pro 0813	DeepSeek	Open	Reasoning, Max Effort	53.2	68.8	not measured	66.9	$0.35 / $0.8	1.05M
23	Muse Spark 1.1	Meta	Closed	xhigh	53.2	71.3	not measured	not measured	$1.25 / $4.25	1.05M
24	GPT-5.4	OpenAI	Closed	xhigh	53.1	71.1	not measured	not measured	$0.75 / $6	1.05M
25	GLM-5.2	Z.AI	Open	max	52.6	68.8	not measured	68.6	$0.3 / $1.05	1M
26	GPT-5.6 Luna	OpenAI	Closed	max	52.3	71.4	not measured	116.2	$0.1 / $0.6	1.05M
27	Gemini 3.5 Flash	Google	Closed	high	52.0	70.1	not measured	250.6	$0.186 / $1.11	1.05M
28	Qwen3.8 27B	Alibaba	Open	xhigh	52.0	68.1	not measured	42.9	$0.1 / $0.4	262K
29	DeepSeek V4 Flash 0731	DeepSeek	Open	Reasoning, Max Effort	51.8	69.1	not measured	129.7	$0.03 / $0.1	1M
30	Gemini 3.6 Flash	Google	Closed	high	51.6	69.2	not measured	187.8	$0.375 / $1.88	1.05M
31	Claude Sonnet 4.6	Anthropic	Closed	Adaptive Reasoning, Max Effort	48.4	63.0	not measured	44.6	$0.9 / $5.55	1M
32	Gemini 3.1 Pro Preview	Google	Closed	default	47.7	68.8	not measured	117.4	$1 / $6	1.05M
33	Motif 3	Motif Technologies	Closed	default	47.4	63.5	not measured	not measured	$0.5 / $2	262K
34	Qwen3.7 Max	Alibaba	Closed	default	46.7	66.0	not measured	not measured	$0.825 / $2.48	1M
35	GPT-5.3 Codex	OpenAI	Closed	xhigh	45.5	not measured	not measured	126.0	$1.6 / $13	4K
36	MiniMax-M3	MiniMax	Open	default	45.4	58.6	not measured	77.1	$0.225 / $0.9	1.05M
37	DeepSeek V4 Pro	DeepSeek	Open	Reasoning, Max Effort	45.3	59.4	not measured	67.8	$0.348 / $0.696	1M
38	Kimi K2.6	Moonshot AI	Open	default	45.1	61.8	not measured	not measured	$0.22 / $1.14	262K
39	Claude Opus 4.6	Anthropic	Closed	Adaptive Reasoning, Max Effort	44.9	not measured	not measured	not measured	$2.5 / $12.5	1M
40	GPT-5.2	OpenAI	Closed	xhigh	43.3	not measured	99.0	not measured	$0.25 / $2	4K
41	Kimi K2.7 Code	Moonshot AI	Open	default	43.0	60.8	not measured	38.1	$0.275 / $1.1	262K
42	MiMo-V2.5-Pro	Xiaomi	Open	default	42.9	60.2	not measured	37.6	$0.4 / $0.8	1.05M
43	Inkling	Thinking Machines	Open	xhigh	42.3	52.1	not measured	70.5	$0.95 / $4.05	1.05M
44	Hy3	Tencent	Open	default	42.2	58.8	not measured	92.5	$0.132 / $0.528	262K
45	DeepSeek V4 Flash	DeepSeek	Open	Reasoning, Max Effort	42.1	56.2	not measured	not measured	$0.051 / $0.104	1M
46	Claude Opus 4.5	Anthropic	Closed	Reasoning	41.9	not measured	91.3	not measured	$0.71 / $3.57	2K
47	Nex N2 Pro	NEX AGI	Open	default	41.7	59.1	not measured	134.0	$0.25 / $1	262K
48	Solar Pro 4	Upstage	Closed	default	41.6	52.7	not measured	53.6	$0.03 / $0.12	524K
49	MiMo-V2-Pro	Xiaomi	Open	default	41.4	not measured	not measured	not measured	$0.435 / $0.87	1.05M
50	GPT-5.2 Codex	OpenAI	Closed	xhigh	41.2	not measured	not measured	not measured	$0.14 / $1.14	4K
51	Inkling Small	Thinking Machines	Open	default	41.2	52.9	not measured	76.9	$0.45 / $1.2	524K
52	Qwen3.6 Max Preview	Alibaba	Open	default	41.1	not measured	not measured	not measured	$1.03 / $6.16	262K
53	GLM-5.1	Z.AI	Open	Reasoning	41.0	55.8	not measured	not measured	$0.45 / $2.15	2K
54	GPT-5.4 mini	OpenAI	Closed	xhigh	40.9	56.1	not measured	not measured	$0.375 / $2.25	4K
55	GLM-5	Z.AI	Open	Reasoning	40.6	not measured	not measured	not measured	$0.6 / $1.92	203K
56	Gemini 3 Pro Preview	Google	Closed	high	40.6	not measured	95.7	not measured	$0.57 / $3.43	1.05M
57	Qwen3.6 Plus	Alibaba	Open	default	40.5	54.5	not measured	not measured	$0.276 / $1.65	1M
58	Agnes 2.5 Pro Alpha	Sapiens AI	Open	default	39.7	58.8	not measured	196.1	$0.45 / $0.9	1M
59	GPT-5.4 nano	OpenAI	Closed	xhigh	39.7	56.1	not measured	not measured	$0.1 / $0.625	4K
60	Qwen3.7 Plus	Alibaba	Closed	default	39.4	55.9	not measured	56.4	$0.282 / $1.13	1M
61	GLM-5-Turbo	Z.AI	Open	default	39.1	not measured	not measured	not measured	$0.72 / $3.2	2K
62	MiniMax-M2.7	MiniMax	Open	default	38.9	52.6	not measured	not measured	$0.08 / $0.32	205K
63	Gemini 3 Flash Preview	Google	Closed	Reasoning	38.7	not measured	97.0	not measured	$0.07 / $0.43	1.05M
64	Nemotron 3 Ultra 550B A55B	NVIDIA	Open	Reasoning	38.3	49.3	not measured	148.3	$0.1 / $0.1	1.05M
65	MiMo-V2.5	Xiaomi	Open	default	38.0	56.8	not measured	58.3	$0.14 / $0.28	1.05M
66	Grok 4.3	xAI	Closed	high	37.9	42.2	not measured	129.3	$1.25 / $2.5	1M
67	Ling 3.0 Flash	InclusionAI	Open	default	37.8	50.6	not measured	336.0	$0.021 / $0.063	262K
68	Qwen3.6 27B	Alibaba	Open	Reasoning	37.7	53.7	not measured	52.7	$0.6 / $0.6	262K
69	GPT-5.1	OpenAI	Closed	high	37.5	49.4	94.0	not measured	$0.625 / $5	4K
70	Claude Sonnet 4.5	Anthropic	Closed	Reasoning	37.4	52.1	88.0	not measured	$0.43 / $2.14	2K
71	Gemini 3.5 Flash Lite	Google	Closed	default	37.4	49.3	not measured	364.4	$0.15 / $1.25	1.05M
72	GPT-5-Codex	OpenAI	Closed	high	37.0	not measured	98.7	not measured	$1.07 / $8.5	4K
73	Kimi K2.5	Moonshot AI	Open	Reasoning	36.0	46.8	not measured	not measured	$0.3 / $1.9	262K
74	MiMo V2 Omni	Xiaomi	Open	default	35.9	not measured	not measured	not measured	$0.14 / $0.28	262K
75	GPT-5.1 Codex	OpenAI	Closed	high	35.6	not measured	95.7	not measured	$1.07 / $8.5	4K
76	GLM-5V-Turbo	Z.AI	Closed	Reasoning	35.3	not measured	not measured	not measured	$0.72 / $3.2	2K
77	GPT-5	OpenAI	Closed	high	35.3	37.8	94.3	not measured	$0.625 / $5	4K
78	muse-glimmer	Meta	Closed	high	35.1	49.0	not measured	108.3	not published / not published	not published
79	Qwen3.5 27B	Alibaba	Open	Reasoning	34.6	not measured	not measured	not measured	$0.086 / $0.688	262K
80	Claude Opus 4.1	Anthropic	Closed	Reasoning	34.5	not measured	80.3	not measured	$7.5 / $37.5	2K
81	GLM-4.7	Z.AI	Open	Reasoning	34.5	45.3	95.0	not measured	$0.2 / $0.8	205K
82	MiniMax-M2.5	MiniMax	Open	default	34.5	not measured	not measured	not measured	$0.22 / $0.88	205K
83	Hy3 preview	Tencent	Open	Reasoning	34.4	not measured	not measured	not measured	$0.066 / $0.26	262K
84	GPT-5.5 Instant	OpenAI	Closed	May 2026	34.3	not measured	not measured	148.0	$5 / $30	4K
85	Qwen3.5 397B-A17B	Alibaba	Open	Reasoning	34.3	48.2	not measured	92.0	$0.172 / $1.03	262K
86	Grok 4	xAI	Closed	default	34.1	not measured	92.7	not measured	$3 / $15	256K
87	LongCat-2.0	Meituan	Open	default	34.0	45.3	not measured	40.5	$0.3 / $1.2	1.05M
88	MiMo-V2-Flash	Xiaomi	Open	Feb 2026	34.0	not measured	not measured	not measured	$0.1 / $0.3	262K
89	Kat Coder Pro V2	Kwaipilot	Closed	default	33.7	59.5	not measured	not measured	$0.3 / $1.2	262K
90	Kimi K2 Thinking	Moonshot AI	Open	default	33.5	not measured	94.7	not measured	$0.47 / $2	262K
91	o3-pro	OpenAI	Closed	default	33.3	not measured	not measured	not measured	$20 / $40	2K
92	DeepSeek V3.2	DeepSeek	Open	Reasoning	32.8	44.2	92.0	not measured	$0.18 / $0.35	164K
93	Qwen3.5 122B-A10B	Alibaba	Open	Reasoning	32.8	45.7	not measured	135.4	$0.115 / $0.917	262K
94	Qwen 3 Max Thinking	Alibaba	Open	default	32.5	not measured	not measured	not measured	$0.78 / $3.9	256K
95	MiniMax-M2.1	MiniMax	Open	default	32.1	not measured	82.7	not measured	$0.27 / $1.1	205K
96	Qwen3.6 35B-A3B	Alibaba	Open	Reasoning	32.1	41.9	not measured	141.4	$0.07 / $0.42	262K
97	Claude Opus 4	Anthropic	Closed	Reasoning	31.7	not measured	73.3	not measured	$13 / $64	2K
98	Ring-2.6-1T	InclusionAI	Open	default	31.7	42.8	not measured	128.9	$0.3 / $2.5	262K
99	GPT-5 Mini	OpenAI	Closed	medium	31.6	not measured	85.0	not measured	$0.04 / $0.29	4K
100	DeepSeek V3.1 Terminus	DeepSeek	Open	Reasoning	31.4	43.5	89.7	not measured	$0.25 / $0.7	128K
101	GPT-5.1 Codex mini	OpenAI	Closed	high	31.3	not measured	91.7	not measured	$0.22 / $1.8	4K
102	Grok 4.1 Fast	xAI	Closed	Reasoning	31.3	not measured	89.3	not measured	$0.2 / $0.5	2M
103	Qwen3.5 Omni Plus	Alibaba	Closed	default	31.3	not measured	not measured	125.8	free / free	984K
104	o3	OpenAI	Closed	default	31.1	not measured	88.3	128.7	$1 / $4	2K
105	Step 3.7 Flash	StepFun	Open	default	30.9	39.6	not measured	101.9	$0.185 / $1.11	256K
106	Mistral Medium 3.5	Mistral	Open	default	30.4	46.9	not measured	124.9	$0.75 / $3.75	262K
107	Claude Haiku 4.5	Anthropic	Closed	Reasoning	29.9	43.9	83.7	100.2	$0.14 / $0.71	2K
108	Qwen3.5 35B-A3B	Alibaba	Open	Reasoning	29.9	not measured	not measured	157.3	$0.057 / $0.459	262K
109	Claude Sonnet 4	Anthropic	Closed	Reasoning	29.8	37.6	74.3	not measured	$2.6 / $13	2K
110	Gemma 4 31B	Google	Open	Reasoning	29.7	43.4	not measured	35.2	$0.102 / $0.297	262K
111	GLM-4.6	Z.AI	Open	Reasoning	29.3	45.8	86.0	not measured	$0.286 / $1.14	205K
112	KAT-Coder-Pro V1	Kwaipilot	Closed	default	28.9	not measured	94.7	not measured	$0.3 / $1.2	256K
113	MiniMax-M2	MiniMax	Open	default	28.9	not measured	78.3	not measured	$0.2 / $1	205K
114	Grok 4 Fast	xAI	Closed	Reasoning	27.9	not measured	89.7	not measured	$0.2 / $0.5	2M
115	Claude 3.7 Sonnet	Anthropic	Closed	Reasoning	27.6	36.4	56.3	not measured	$2.6 / $13	2K
116	Ling-2.6-1T	InclusionAI	Open	default	26.6	not measured	not measured	not measured	$0.3 / $2.5	262K
117	Step 3.5 Flash 2603	StepFun	Open	default	26.5	not measured	not measured	not measured	$0.1 / $0.3	256K
118	Gemma 4 26B A4B	Google	Open	Reasoning	26.1	39.3	not measured	68.8	$0.042 / $0.22	262K
119	o4-mini	OpenAI	Closed	high	26.1	not measured	90.7	not measured	$0.55 / $2.2	2K
120	Step 3.5 Flash	StepFun	Open	default	26.0	not measured	not measured	not measured	$0.09 / $0.3	256K
121	DeepSeek V3.2 Exp	DeepSeek	Open	Reasoning	25.9	not measured	87.7	not measured	$0.22 / $0.33	164K
122	Gemini 2.5 Pro	Google	Closed	default	25.9	33.3	87.7	not measured	$0.625 / $5	1.05M
123	Nemotron 3 Super 120B A12B	NVIDIA	Open	Reasoning	25.7	37.7	not measured	141.6	$0.085 / $0.4	262K
124	Gemini 3.1 Flash Lite	Google	Closed	default	25.6	34.7	not measured	not measured	$0.125 / $0.75	1.05M
125	Qwen3 Max	Alibaba	Closed	default	24.5	not measured	80.7	not measured	$0.36 / $1.43	262K
126	ling-3.0-tiny	InclusionAI	Closed	default	24.5	26.5	not measured	206.2	free / free	262K
127	Gemini 2.5 Flash Preview	Google	Closed	Sep '25, Reasoning	24.2	not measured	78.3	not measured	$0.15 / $0.6	1.05M
128	GPT OSS 120B	OpenAI	Open	high	24.1	30.4	93.4	151.6	$0.039 / $0.1	131K
129	o1	OpenAI	Closed	default	23.9	39.7	not measured	not measured	$14 / $54	2K
130	Nemotron 3.5 Lightning	NVIDIA	Open	default	23.6	26.8	not measured	283.5	$0.05 / $0.2	1M
131	GLM-4.7-Flash	Z.AI	Open	Reasoning	23.3	not measured	not measured	not measured	$0.06 / $0.4	2K
132	Command A Plus	Cohere	Open	default	22.8	27.8	not measured	259.1	$2.5 / $10	128K
133	DeepSeek-V3.2-Speciale	DeepSeek	Open	default	22.6	not measured	96.7	not measured	$0.58 / $1.68	128K
134	Ernie 5.0 Thinking Preview	Baidu	Closed	default	22.3	not measured	85.0	not measured	$1 / $3.5	128K
135	Gemma 4 12B Instruct	Google	Open	Reasoning	22.2	31.0	not measured	131.1	$0.05 / $0.25	262K
136	Grok Code Fast 1	xAI	Closed	default	22.0	not measured	43.3	not measured	$0.18 / $1.35	256K
137	Mercury 2	Inception	Closed	default	21.9	31.1	not measured	854.3	$0.25 / $0.75	128K
138	Qwen3.5 9B	Alibaba	Open	Reasoning	21.8	28.7	not measured	84.6	$0.04 / $0.15	262K
139	DeepSeek V3.1	DeepSeek	Open	Non-reasoning	21.4	not measured	49.7	not measured	$0.2 / $0.7	164K
140	Qwen3 Coder Next	Alibaba	Open	default	21.3	36.2	not measured	111.6	$0.108 / $0.675	262K
141	Qwen3 VL 235B A22B	Alibaba	Open	Reasoning	20.9	not measured	88.3	not measured	$0.2 / $0.88	131K
142	DeepSeek R1 0528	DeepSeek	Open	May '25	20.4	not measured	76.0	not measured	$0.4 / $1.7	164K
143	Qwen3.5 4B	Alibaba	Open	Reasoning	20.4	22.6	not measured	21.0	$0.04 / $0.07	262K
144	Gemini 2.5 Flash	Google	Closed	Reasoning	20.3	not measured	73.3	not measured	$0.09 / $0.71	1.05M
145	North Mini Code	Cohere	Open	default	20.2	36.5	not measured	81.4	free / free	256K
146	GPT-5 Nano	OpenAI	Closed	high	20.1	not measured	83.7	not measured	$0.025 / $0.2	4K
147	Qwen 3 235b A22B 2507	Alibaba	Open	Reasoning	19.9	22.1	91.0	not measured	$0.072 / $0.464	262K
148	GLM-4.5	Z.AI	Open	Reasoning	19.7	not measured	73.7	not measured	$0.286 / $1.14	131K
149	Kimi K2	Moonshot AI	Open	default	19.7	not measured	57.0	not measured	$0.4 / $1.8	131K
150	Mistral Small 4	Mistral	Open	Reasoning	19.7	26.6	not measured	152.8	$0.143 / $0.568	256K
151	GPT-4.1	OpenAI	Closed	default	19.6	not measured	34.7	not measured	$1 / $4	1.05M
152	Granite 4.2 8B	IBM	Open	default	19.6	22.4	not measured	148.7	$0.1 / $0.15	131K
153	Devstral 2	Mistral	Open	default	19.2	31.3	36.7	126.1	$0.4 / $2	262K
154	Qwen3.5 Omni Flash	Alibaba	Closed	default	19.2	not measured	not measured	224.3	free / free	49K
155	o3-mini	OpenAI	Closed	default	19.2	not measured	not measured	not measured	$0.55 / $2.2	2K
156	o1-pro	OpenAI	Closed	default	19.1	not measured	not measured	not measured	$140 / $540	2K
157	Trinity Large Thinking	Arcee AI	Open	default	18.7	25.8	not measured	320.4	$0.25 / $0.8	262K
158	DeepSeek-R1	DeepSeek	Open	Jan '25	18.6	24.6	68.0	not measured	$0.4 / $1.7	128K
159	Grok 3	xAI	Closed	default	18.6	not measured	58.0	not measured	$3 / $15	131K
160	Qwen3 235B A22B Instruct 2507	Alibaba	Open	default	18.4	not measured	71.7	not measured	$0.087 / $0.35	262K
161	Qwen3 Coder 480B A35B Instruct	Alibaba	Open	default	18.2	not measured	39.3	not measured	$0.38 / $1.55	262K
162	Sonar Reasoning Pro	Perplexity	Closed	default	18.0	not measured	not measured	not measured	$2 / $8	128K
163	MiniMax M1 80K	MiniMax	Closed	default	17.9	not measured	61.0	not measured	$0.605 / $2.42	1M
164	Devstral Small 2	Mistral	Open	default	17.7	29.3	34.3	124.8	$0.1 / $0.3	256K
165	GLM-4.6V	Z.AI	Open	Reasoning	16.9	not measured	85.3	not measured	$0.14 / $0.42	131K
166	Qwen3-Next 80B-A3B (Thinking)	Alibaba	Open	Reasoning	16.9	17.4	84.3	188.5	$0.15 / $0.65	131K
167	GLM-4.5-Air	Z.AI	Open	default	16.7	not measured	80.7	not measured	$0.114 / $0.286	131K
168	Ring-1T	InclusionAI	Open	default	16.3	not measured	89.3	not measured	$0.56 / $2.24	128K
169	Mistral Large 3	Mistral	Open	default	15.9	20.1	38.0	72.6	$0.5 / $1.5	262K
170	intellect-3	Prime Intellect	Closed	default	15.7	not measured	88.0	not measured	not published / not published	not published
171	DeepSeek V3 0324	DeepSeek	Open	default	15.2	21.2	41.0	not measured	$0.2 / $0.8	164K
172	GPT OSS 20B	OpenAI	Open	high	15.2	20.7	89.3	127.7	$0.02 / $0.1	131K
173	Gemini 2.5 Flash Lite Preview	Google	Closed	Sep '25, Reasoning	15.2	not measured	68.7	not measured	$0.15 / $0.6	1.05M
174	Nemotron 3 Nano Omni 30B A3B Reasoning	NVIDIA	Open	default	15.0	13.8	not measured	329.0	$0.2 / $0.8	262K
175	Mistral Small 3.1	Mistral	Closed	default	14.9	26.3	3.7	not measured	$0.1 / $0.3	128K
176	GPT-4.1 mini	OpenAI	Closed	default	14.8	20.2	46.3	not measured	$0.2 / $0.8	1.05M
177	Mistral Medium 3.1	Mistral	Closed	default	14.7	20.5	38.3	not measured	$0.4 / $2	131K
178	Qwen3 30B A3B 2507	Alibaba	Open	Reasoning	14.6	12.1	56.3	not measured	free / free	262K
179	Llama 4 Maverick	Meta	Open	default	14.5	16.3	19.3	82.9	$0.15 / $0.6	1.05M
180	Nemotron 3 Nano 30B A3B	NVIDIA	Open	Reasoning	14.5	14.4	91.0	219.3	$0.05 / $0.2	262K
181	Solar Pro 3	Upstage	Closed	default	14.5	16.2	not measured	145.0	$0.25 / $0.25	131K
182	Qwen3 VL 235B A22B Instruct	Alibaba	Open	default	14.4	not measured	70.7	not measured	$0.2 / $0.88	262K
183	DeepSeek V3	DeepSeek	Open	Dec '24	14.2	23.0	26.0	not measured	$0.32 / $0.89	164K
184	Ling-2.6-flash	InclusionAI	Open	default	14.2	25.3	not measured	not measured	$0.1 / $0.3	262K
185	Qwen3-Next 80B-A3B Instruct	Alibaba	Open	default	13.8	not measured	66.3	184.0	$0.144 / $0.574	131K
186	Qwen3-Coder 30B-A3B Instruct	Alibaba	Open	default	13.6	not measured	29.0	not measured	$0.06 / $0.25	262K
187	DiffusionGemma 26B-A4B IT	Google	Closed	default	13.5	19.7	not measured	not measured	$0.5 / $0.5	262K
188	Qwen3 235B-A22B	Alibaba	Open	Reasoning	13.5	not measured	82.0	not measured	$0.13 / $0.6	131K
189	QwQ 32B	Alibaba	Open	default	13.4	not measured	29.0	not measured	$0.4 / $0.4	131K
190	Qwen3-VL 30B-A3B	Alibaba	Open	Reasoning	13.4	not measured	82.3	not measured	$0.108 / $0.431	262K
191	Ling-1T	InclusionAI	Open	default	12.7	not measured	71.3	not measured	$0.56 / $2.24	128K
192	Mistral Medium 3	Mistral	Closed	default	12.5	not measured	30.3	not measured	$0.4 / $2	131K
193	Solar Pro 2	Upstage	Closed	Preview, Reasoning	12.5	not measured	not measured	not measured	$0.25 / $0.25	66K
194	Celeris 1	Celeris	Closed	default	12.4	14.4	not measured	2046.9	$2 / $6	8K
195	GPT-4o	OpenAI	Closed	March 2025, chatgpt-4o-latest	12.3	not measured	25.7	not measured	$1.25 / $5	128K
196	Claude 3.5 Haiku	Anthropic	Closed	default	12.2	15.9	not measured	not measured	$0.68 / $3.4	2K
197	Gemini 2.0 Flash	Google	Closed	Feb '25	12.2	not measured	21.7	not measured	$0.1 / $0.42	1.05M
198	Gemma 4 E4B Instruct	Google	Open	Reasoning	12.2	9.4	not measured	34.1	$0.04 / $0.2	33K
199	Llama 3.3 Nemotron Super 49B v1	NVIDIA	Open	Reasoning	12.2	not measured	54.7	not measured	free / free	131K
200	MiniCPM5-1B	OpenBMB	Open	Reasoning	11.9	not measured	not measured	not measured	$0.124 / $0.743	131K
201	Sarvam 105B	Sarvam	Open	high	11.9	not measured	not measured	not measured	$0.04 / $0.16	131K
202	Magistral Small 1.2	Mistral	Open	default	11.5	14.7	80.3	not measured	$0.5 / $1.5	128K
203	Gemini 2.5 Flash-Lite	Google	Closed	Reasoning	11.4	not measured	53.3	not measured	$0.1 / $0.1	1.05M
204	Qwen3 32B	Alibaba	Open	Reasoning	11.4	15.3	73.0	not measured	$0.09 / $0.25	131K
205	Ministral 3 14B	Mistral	Open	default	11.2	14.4	30.0	97.0	$0.2 / $0.2	262K
206	DeepSeek R1 Distill Qwen 32B	DeepSeek	Open	default	11.0	not measured	63.0	not measured	$0.3 / $0.3	8K
207	LFM2.5 2.6B	Liquid AI	Open	default	11.0	7.7	not measured	1914.3	$0.1 / $0.2	128K
208	Qwen3 VL 32B Instruct	Alibaba	Open	default	11.0	not measured	68.3	not measured	$0.104 / $0.416	131K
209	Mistral Small 3.2	Mistral	Open	default	10.7	12.5	27.0	not measured	$0.075 / $0.2	128K
210	Qwen3 14B	Alibaba	Open	Reasoning	10.4	13.8	55.7	not measured	$0.05 / $0.22	131K
211	DeepSeek R1 0528 Qwen3 8B	DeepSeek	Open	default	10.3	not measured	63.7	not measured	$0.06 / $0.09	128K
212	Llama 4 Scout	Meta	Open	default	10.3	8.2	14.0	98.5	$0.1 / $0.3	1.31M
213	Qwen 2.5 Max	Alibaba	Closed	default	10.1	not measured	not measured	not measured	$1.6 / $6.39	32K
214	Qwen3 VL 30B A3B Instruct	Alibaba	Closed	default	9.9	not measured	72.3	not measured	$0.15 / $0.6	262K
215	Claude 3.5 Sonnet	Anthropic	Closed	Oct '24	9.8	30.2	not measured	not measured	$2.6 / $13	2K
216	DeepSeek R1 Distill Llama 70B	DeepSeek	Open	default	9.8	not measured	53.7	not measured	$0.03 / $0.13	131K
217	DeepSeek R1 Distill Qwen 14B	DeepSeek	Open	default	9.7	not measured	55.7	not measured	$0.15 / $0.15	33K
218	GPT-4.1 nano	OpenAI	Closed	default	9.6	11.1	24.0	not measured	$0.05 / $0.2	1.05M
219	Ling-flash-2.0	InclusionAI	Closed	default	9.6	not measured	65.3	not measured	$0.14 / $0.57	131K
220	Gemma 4 E2B Instruct	Google	Open	Reasoning	9.5	7.2	not measured	not measured	$0.02 / $0.1	131K
221	Qwen2.5 72B	Alibaba	Open	default	9.4	not measured	14.0	not measured	$0.11 / $0.38	131K
222	Sonar	Perplexity	Closed	default	9.4	not measured	not measured	not measured	$1 / $1	128K
223	Llama 3.3 70B	Meta	Open	default	9.3	11.9	7.7	85.6	$0.05 / $0.23	128K
224	Qwen3 30B A3B	Alibaba	Open	Reasoning	9.2	not measured	72.3	not measured	$0.08 / $0.29	41K
225	Sonar Pro	Perplexity	Closed	default	9.1	not measured	not measured	not measured	$3 / $15	2K
226	GLM-4.5V	Z.AI	Open	Reasoning	9.0	not measured	73.0	not measured	$0.29 / $0.86	66K
227	Ministral 3 8B	Mistral	Open	default	9.0	9.7	31.7	113.5	$0.15 / $0.15	262K
228	ERNIE 4.5 300B A47B	Baidu	Open	default	8.9	not measured	41.3	not measured	$0.28 / $1.1	131K
229	Qwen3 30B A3B Instruct 2507	Alibaba	Open	default	8.9	not measured	66.3	not measured	$0.048 / $0.193	262K
230	llama-3.1-nemotron-ultra-253b-v1	NVIDIA	Closed	Reasoning	8.9	not measured	63.7	52.9	$0.598 / $1.79	128K
231	Nemotron Nano 12B v2 VL	NVIDIA	Open	Reasoning	8.8	not measured	75.0	39.9	$0.2 / $0.6	128K
232	Nemotron Nano 9B V2	NVIDIA	Open	Reasoning	8.7	not measured	69.7	100.5	$0.06 / $0.23	131K
233	Gemini 2.0 Flash Lite	Google	Closed	Feb '25	8.6	not measured	not measured	not measured	$0.052 / $0.21	2M
234	Llama 3.1 405B Instruct	Meta	Open	default	8.3	not measured	3.0	not measured	$1.95 / $1.95	128K
235	Qwen3 8B	Alibaba	Open	Reasoning	8.3	9.0	19.0	not measured	$0.035 / $0.138	131K
236	Qwen3 4B	Alibaba	Open	Reasoning	8.2	not measured	22.3	not measured	$0.03 / $0.03	128K
237	Qwen3 VL 8B Instruct	Alibaba	Open	default	8.2	not measured	27.3	not measured	$0.117 / $0.455	262K
238	Pixtral Large (25.02)	Mistral	Closed	default	8.0	not measured	2.3	not measured	$1.99 / $5.98	128K
239	ring-flash-2.0	InclusionAI	Closed	default	8.0	not measured	83.7	not measured	not published / not published	not published
240	GPT-4 Turbo	OpenAI	Closed	default	7.7	21.5	not measured	not measured	$5 / $15	128K
241	Command A	Cohere	Open	default	7.5	not measured	13.0	56.5	$2.5 / $10	256K
242	Nova Pro	Amazon	Closed	default	7.5	not measured	7.0	not measured	$0.56 / $2.13	3K
243	Gemma 3 27B	Google	Open	default	7.4	10.1	20.7	not measured	$0.08 / $0.16	131K
244	Llama 3.1 8B	Meta	Open	default	7.4	5.4	4.3	not measured	$0.025 / $0.025	131K
245	Llama 3.1 Nemotron 70B Instruct	NVIDIA	Open	default	7.4	not measured	11.0	34.5	$0.6 / $0.6	131K
246	Qwen3.5 2B	Alibaba	Open	Reasoning	7.4	2.9	not measured	not measured	$0.02 / $0.1	262K
247	Qwen2.5 32B Instruct	Alibaba	Open	default	7.2	not measured	not measured	not measured	$0.287 / $0.861	131K
248	Ministral 3 3B	Mistral	Open	default	7.1	4.8	22.0	208.1	$0.1 / $0.1	256K
249	Qwen2.5 Coder 32B	Alibaba	Open	default	6.9	not measured	not measured	not measured	$0.06 / $0.2	33K
250	GPT-4	OpenAI	Closed	default	6.8	13.1	not measured	not measured	$30 / $60	8K
251	GPT-4o mini	OpenAI	Closed	default	6.7	11.4	14.7	not measured	$0.075 / $0.3	128K
252	Mistral Small 3	Mistral	Open	default	6.7	not measured	4.3	not measured	$0.05 / $0.08	33K
253	Nova Lite	Amazon	Closed	default	6.7	not measured	7.0	not measured	$0.06 / $0.24	3K
254	Llama 3.1 70B	Meta	Open	default	6.5	not measured	4.0	not measured	$0.4 / $0.4	128K
255	Granite 4.1 8B	IBM	Open	default	6.4	9.5	not measured	105.7	$0.05 / $0.1	131K
256	Sarvam 30B	Sarvam	Open	high	6.4	not measured	not measured	not measured	$0.02 / $0.1	128K
257	DeepSeek R1 Distill Llama 8B	DeepSeek	Closed	default	6.2	not measured	41.3	not measured	free / free	33K
258	Saba	Mistral	Open	default	6.2	not measured	not measured	not measured	$0.199 / $0.595	33K
259	solar-mini	Upstage	Closed	default	6.0	not measured	not measured	not measured	$0.15 / $0.15	33K
260	Phi-4-mini	Microsoft	Open	default	5.7	3.8	6.7	46.5	$0.075 / $0.3	128K
261	Gemma 3 12B	Google	Open	default	5.5	5.8	18.3	not measured	$0.05 / $0.1	131K
262	Qwen3.5 0.8B	Alibaba	Open	Reasoning	5.2	0.0	not measured	not measured	$0.06 / $0.12	262K
263	Granite-4.0-H-Small	IBM	Open	default	4.9	not measured	13.7	41.6	$0.064 / $0.265	131K
264	Qwen3 Omni 30B A3B Instruct	Alibaba	Open	default	4.8	not measured	52.3	105.2	$0.25 / $0.97	66K
265	LFM2 24B A2B	Liquid AI	Open	default	4.6	not measured	not measured	not measured	$0.03 / $0.12	33K
266	Phi 4	Microsoft	Open	default	4.6	not measured	18.0	40.7	$0.07 / $0.14	128K
267	Nova Micro	Amazon	Closed	default	4.4	not measured	6.0	248.9	$0.03 / $0.1	128K
268	Mistral Small (latest)	Mistral	Open	Sep '24	4.3	not measured	not measured	not measured	$0.1 / $0.3	256K
269	Phi 4 Multimodal	Microsoft	Open	default	4.2	not measured	not measured	24.5	$0.07 / $0.11	128K
270	Mistral Large	Mistral	Open	Feb '24	4.1	not measured	not measured	not measured	$0.5 / $1.5	262K
271	Qwen2.5-Coder 7B Instruct	Alibaba	Open	default	4.1	not measured	not measured	not measured	$0.144 / $0.287	131K
272	Mixtral 8x22B	Mistral	Open	default	4.0	not measured	not measured	not measured	$2 / $6	66K
273	Llama 3.2 3B	Meta	Open	default	3.9	not measured	3.3	not measured	$0.02 / $0.02	131K
274	Reka Flash 3	Reka	Open	default	3.7	not measured	33.7	90.2	$0.1 / $0.2	66K
275	Claude 3 Haiku	Anthropic	Closed	default	3.5	not measured	not measured	not measured	$0.21 / $1.1	2K
276	DeepSeek R1 Distill Qwen 1.5B	DeepSeek	Closed	default	3.3	not measured	22.0	not measured	free / free	33K
277	GPT-3.5-turbo	OpenAI	Closed	default	3.2	10.7	not measured	not measured	$0.25 / $0.75	16K
278	Mistral Medium (latest)	Mistral	Open	default	3.2	not measured	not measured	not measured	$0.4 / $2	262K
279	Llama 3 70B Instruct	Meta	Open	default	3.1	not measured	not measured	not measured	$0.51 / $0.74	8K
280	Llama 3.2 11B Instruct	Meta	Open	Vision	3.0	not measured	1.7	24.0	$0.07 / $0.33	128K
281	Command R+	Cohere	Open	Apr '24	2.6	not measured	not measured	not measured	$2.5 / $10	128K
282	sarvam-m	Sarvam AI	Open	Reasoning	2.6	not measured	not measured	not measured	free / free	128K
283	Apertus 70B	Swiss AI Initiative	Open	default	2.0	not measured	not measured	not measured	$0.46 / $2.42	66K
284	Granite 4.0 Micro	IBM	Open	default	2.0	not measured	6.0	not measured	$0.017 / $0.112	131K
285	Mixtral 8x7B	Mistral	Open	default	2.0	not measured	not measured	not measured	$0.7 / $0.7	32K
286	Command R	Cohere	Open	Mar '24	1.7	not measured	not measured	not measured	$0.15 / $0.6	128K
287	Apertus 8B Instruct	Swiss AI Initiative	Open	default	1.0	not measured	not measured	not measured	$0.1 / $0.2	66K
288	Gemma 3 4B	Google	Open	default	1.0	2.7	12.7	not measured	$0.04 / $0.08	131K
289	Gemma 3N E4B Instruct	Google	Open	default	1.0	3.2	14.3	not measured	$0.06 / $0.12	33K
290	Gemma 3n E2b It	Google	Open	default	1.0	not measured	10.3	not measured	free / free	128K
291	Llama 3 8B Instruct	Meta	Open	default	1.0	not measured	not measured	not measured	$0.04 / $0.04	8K
292	Llama 3.2 1B	Meta	Open	default	1.0	not measured	0.0	not measured	$0.01 / $0.01	6K

How to read these LLM benchmarks

The Artificial Analysis Intelligence, Coding and Math indices are reported on a 0–100 scale. A model can have several evaluated reasoning settings; the leaderboard uses its highest Intelligence configuration and shows the speed measured for that same run when available. Missing measurements are never treated as zero.

Read the full methodology and source limitations.