What is the Artificial Analysis Intelligence Index?

The Artificial Analysis Intelligence Index v4.0 is a composite benchmark that aggregates performance across 10 evaluations spanning mathematics, science, coding, agentic tasks, and reasoning to measure AI capabilities holistically.

How is the Intelligence Index calculated?

The index aggregates scores from 10 benchmarks: GDPval-AA, τ²-Bench, Terminal-Bench Hard, SciCode, AA-LCR, AA-Omniscience, IFBench, Humanity's Last Exam, GPQA Diamond, and CritPt. All tests are independently run on standardized hardware.

How does this differ from LMArena?

LMArena uses crowdsourced user votes (Elo ratings) reflecting subjective preferences. The AA Intelligence Index uses standardized automated benchmarks with objective scoring across specific technical domains.

Where can I find the original data?

The original leaderboard is available at artificialanalysis.ai/leaderboards/models and the methodology at artificialanalysis.ai/evaluations/artificial-analysis-intelligence-index.

Artificial Analysis Intelligence Index AI模型智能指数排行榜

Name: Artificial Analysis Intelligence Index AI模型智能指数排行榜
Creator: DataLearner
License: https://creativecommons.org/licenses/by/4.0/

Artificial Analysis Intelligence Index v4.0 综合了10项权威评测基准（GDPval-AA、Terminal-Bench、GPQA Diamond、SciCode等），从数学、科学、编程、推理等多维度对AI模型进行全面评估和排名。

榜首模型

Claude Opus 5 (max)

最高得分

模型数量

242

数据版本

2026年08月08日

数据来源: Artificial Analysis

来源：全部国产模型

榜单历史快照月份:

排名总表

排名	模型名称	智能指数	机构
	Claude Opus 5 (max)Anthropic	63	Anthropic
	Claude Opus 5 (xhigh)Anthropic	63	Anthropic
	Claude Fable 5Anthropic	62	Anthropic
4	Claude Opus 5 (high)Anthropic	61	Anthropic
5	GPT-5.6 Sol (max)OpenAI	61	OpenAI
6	Kimi K3 (max)Moonshot AI	60	Moonshot AI
7	GPT-5.6 Sol (xhigh)OpenAI	59	OpenAI
8	Claude Opus 5 (medium)Anthropic	59	Anthropic
9	Qwen3.8-Max阿里巴巴	58	阿里巴巴
10	GPT-5.6 Sol (high)OpenAI	57	OpenAI
11	Muse Spark 1.2 (xhigh)Meta	57	Meta
12	GPT-5.6 Terra (max)OpenAI	57	OpenAI
13	Grok 4.5 (high)xAI	56	xAI
14	GPT-5.6 Sol (medium)OpenAI	56	OpenAI
15	Claude Sonnet 5 (max)Anthropic	55	Anthropic
16	GPT-5.6 Terra (xhigh)OpenAI	53	OpenAI
17	GLM-5.2 (max)智谱AI	53	智谱AI
18	Claude Opus 5 (low)Anthropic	52	Anthropic
19	GPT-5.6 Luna (max)OpenAI	52	OpenAI
20	DeepSeek-V4-Flash (max)DeepSeek-AI	52	DeepSeek-AI
21	Gemini 3.6 FlashGoogle Deep Mind	52	Google Deep Mind
22	GPT-5.6 Sol (low)OpenAI	51	OpenAI
23	GPT-5.6 Terra (high)OpenAI	50	OpenAI
24	GPT-5.6 Luna (xhigh)OpenAI	50	OpenAI
25	Kimi K3 (low)Moonshot AI	48	Moonshot AI
26	Gemini 3.1 Pro PreviewGoogle Deep Mind	48	Google Deep Mind
27	GPT-5.6 Luna (high)OpenAI	47	OpenAI
28	GPT-5.6 Terra (medium)OpenAI	47	OpenAI
29	Gemini 3.5 Flash (medium)Google Deep Mind	47	Google Deep Mind
30	GPT-5.3 Codex (xhigh)OpenAI	46	OpenAI
31	MiniMax M3MiniMaxAI	45	MiniMaxAI
32	DeepSeek-V4-Pro (max)DeepSeek-AI	45	DeepSeek-AI
33	Motif 3 (Beta)Motif Technologies	45	Motif Technologies
34	DeepSeek-V4-Pro (high)DeepSeek-AI	44	DeepSeek-AI
35	Kimi K2.7 CodeMoonshot AI	43	Moonshot AI
36	MiMo-V2.5-ProXiaomi	43	Xiaomi
37	Claude Sonnet 5 (non-reasoning)Anthropic	43	Anthropic
38	InklingThinking Machines Lab	42	Thinking Machines Lab
39	Hy3腾讯AI实验室	42	腾讯AI实验室
40	GPT-5.6 Sol (non-reasoning)OpenAI	42	OpenAI
41	Nex-N2-ProNex AGI	42	Nex AGI
42	GPT-5.6 Terra (low)OpenAI	41	OpenAI
43	Inkling SmallThinking Machines	41	Thinking Machines
44	JT-4.1 Flash 236B A21BChina Mobile	40	China Mobile
45	Agnes 2.5 Pro AlphaSapiens AI	40	Sapiens AI
46	Qwen3.7-Plus阿里巴巴	39	阿里巴巴
47	GPT-5.6 Luna (medium)OpenAI	39	OpenAI
48	Nemotron 3 UltraNVIDIA	38	NVIDIA
49	MiMo-V2.5Xiaomi	38	Xiaomi
50	Ling 3.0 FlashInclusionAI	38	InclusionAI
51	Qwen3.6-27B阿里巴巴	38	阿里巴巴
52	Gemini 3.5 Flash-LiteGoogle Deep Mind	37	Google Deep Mind
53	Nova 2 Omni（Preview）亚马逊	37	亚马逊
54	Grok 4.3 Beta (medium)xAI	37	xAI
55	Grok 4.3 Beta (low)xAI	36	xAI
56	MiMo-V2-OmniXiaomi	36	Xiaomi
57	Gemini 3.5 Flash (minimal)Google Deep Mind	36	Google Deep Mind
58	Claude Sonnet 4.6 (Non-reasoning, Low Effort)Anthropic	35	Anthropic
59	GLM-5.2智谱AI	35	智谱AI
60	GPT-5.6 Terra (non-reasoning)OpenAI	35	OpenAI
61	Qwen3.5-397B-A17B阿里巴巴	34	阿里巴巴
62	Gemini 2.0 Flash ExperimentalDeepMind	34	DeepMind
63	LongCat 2.0LongCat	34	LongCat
64	KAT-Coder-Pro V2KwaiKAT	34	KwaiKAT
65	GPT-5.6 Luna (low)OpenAI	34	OpenAI
66	Qwen3.5-122B-A10B阿里巴巴	33	阿里巴巴
67	Qwen3.5-397B-A17B阿里巴巴	33	阿里巴巴
68	Qwen3.6-35B-A3B阿里巴巴	32	阿里巴巴
69	DeepSeek-V4-ProDeepSeek-AI	32	DeepSeek-AI
70	G9v3-39A5BAI9Stars	31	AI9Stars
71	Qwen3.5-Omni-Plus阿里巴巴	31	阿里巴巴
72	Qwen3.6-27B阿里巴巴	31	阿里巴巴
73	Ring-2.6-1TInclusionAI	31	InclusionAI
74	OpenAI o3OpenAI	31	OpenAI
75	Step 3.7 FlashStepFunAI	31	StepFunAI
76	Mistral Medium 3.5MistralAI	30	MistralAI
77	Haiku 4.5Anthropic	30	Anthropic
78	Gemma 4 31BDeepMind	30	DeepMind
79	DeepSeek-V4-FlashDeepSeek-AI	29	DeepSeek-AI
80	GPT-5.5 InstantOpenAI	29	OpenAI
81	JT-35B-FlashChina Mobile	29	China Mobile
82	MiMo-V2.5-ProXiaomi	28	Xiaomi
83	Qwen3.5-122B-A10B阿里巴巴	28	阿里巴巴
84	GPT-5.6 Luna (non-reasoning)OpenAI	27	OpenAI
85	Doubao Seed CodeByteDance Seed	26	ByteDance Seed
86	Gemma 4 26B A4BDeepMind	26	DeepMind
87	Nemotron 3 SuperNVIDIA	26	NVIDIA
88	Grok 4.3 Beta (non-reasoning)xAI	25	xAI
89	MiMo-V2-FlashXiaomi	25	Xiaomi
90	Qwen3.6-35B-A3B阿里巴巴	25	阿里巴巴
91	Ling 3.0 TinyInclusionAI	25	InclusionAI
92	Qwen3.5-35B-A3B阿里巴巴	24	阿里巴巴
93	Haiku 4.5Anthropic	24	Anthropic
94	GPT OSS 120B (high)OpenAI	24	OpenAI
95	C4AI Command A (202503)CohereAI	23	CohereAI
96	K-EXAONELG AI Research	22	LG AI Research
97	ERNIE 5.0 Thinking Preview百度	22	百度
98	Gemma 4 12BGoogle	22	Google
99	Gemma 4 31BDeepMind	22	DeepMind
100	Nova 2 Pro（Preview） (medium)亚马逊	22	亚马逊
101	Mercury 2Inception	22	Inception
102	Qwen3.5-9B阿里巴巴	22	阿里巴巴
103	Qwen3-Coder-Next阿里巴巴	21	阿里巴巴
104	Nova 2 Omni（Preview） (medium)亚马逊	21	亚马逊
105	Apriel-v1.6-15B-ThinkerServiceNow	21	ServiceNow
106	Qwen3.5-9B阿里巴巴	21	阿里巴巴
107	EXAONE 4.5 33BLG AI Research	21	LG AI Research
108	Gemma 4 26B A4BDeepMind	20	DeepMind
109	Qwen3.5 4BAlibaba	20	Alibaba
110	North Mini CodeCohere	20	Cohere
111	Nova 2 Pro（Preview） (low)亚马逊	20	亚马逊
112	Mistral Small 4Mistral	20	Mistral
113	Devstral 2Mistral	19	Mistral
114	Nova 2 Lite (medium)亚马逊	19	亚马逊
115	Qwen3.5-Omni-Flash阿里巴巴	19	阿里巴巴
116	JT-MINIChina Mobile	19	China Mobile
117	Trinity Large ThinkingArcee AI	19	Arcee AI
118	Nova 2 Lite (high)亚马逊	18	亚马逊
119	HyperNova 60B 2605Multiverse Computing	18	Multiverse Computing
120	Magistral Medium 1.2Mistral	18	Mistral
121	Nova 2 Lite (low)亚马逊	18	亚马逊
122	Nemotron Cascade 2 30B A3BNVIDIA	18	NVIDIA
123	Devstral Small 2Mistral	18	Mistral
124	LongCat Flash LiteLongCat	17	LongCat
125	K2 Think V2MBZUAI	17	MBZUAI
126	HyperCLOVA X SEED Think (32B)Naver	17	Naver
127	K-EXAONELG AI Research	17	LG AI Research
128	Qwen3-Next阿里巴巴	17	阿里巴巴
129	Nova 2 Omni（Preview） (low)亚马逊	17	亚马逊
130	Mi:dm K 2.5 ProKorea Telecom	17	Korea Telecom
131	G9v3-3BAI9Stars	16	AI9Stars
132	Qwen3.5 4BAlibaba	16	Alibaba
133	Mistral Large 3MistralAI	16	MistralAI
134	INTELLECT-3Prime Intellect	16	Prime Intellect
135	Solar Open 100BUpstage	15	Upstage
136	GPT OSS 20B (high)OpenAI	15	OpenAI
137	Qwen3-Omni-30B-A3B (reasoning)阿里巴巴	15	阿里巴巴
138	GPT OSS 120B (low)OpenAI	15	OpenAI
139	NVIDIA Nemotron 3 NanoNVIDIA	15	NVIDIA
140	Solar Pro 3Upstage	14	Upstage
141	Llama 4 MaverickFacebook AI研究实验室	14	Facebook AI研究实验室
142	Nova 2 Pro（Preview）亚马逊	14	亚马逊
143	GPT OSS 20B (low)OpenAI	14	OpenAI
144	K2-V2 (high)MBZUAI	14	MBZUAI
145	DiffusionGemma 26B A4BGoogle	14	Google
146	Qwen3-Next阿里巴巴	14	阿里巴巴
147	Gemma 4 12B (Non-reasoning)Google	13	Google
148	Motif-2-12.7BMotif Technologies	13	Motif Technologies
149	Nova PremierAmazon	13	Amazon
150	Celeris-1Celeris	12	Celeris
151	K2-V2 (medium)MBZUAI	12	MBZUAI
152	Llama Nemotron Super 49B v1.5Meta	12	Meta
153	Gemma 4 E4BDeepMind	12	DeepMind
154	Mistral Small 4Mistral	12	Mistral
155	Tri-21B-ThinkTrillion Labs	12	Trillion Labs
156	MiniCPM5-1BOpenBMB	12	OpenBMB
157	Sarvam 105B (high)Sarvam	12	Sarvam
158	Nova 2 Lite亚马逊	12	亚马逊
159	MiniCPM5-1BOpenBMB	12	OpenBMB
160	Magistral Small 1.2Mistral	11	Mistral
161	Ministral 3 14BMistralAI	11	MistralAI
162	Nanbeige4.1-3BNanbeige	11	Nanbeige
163	EXAONE 4.0 32BLG AI Research	11	LG AI Research
164	Nova 2 Omni（Preview）亚马逊	10	亚马逊
165	Llama 4 ScoutFacebook AI研究实验室	10	Facebook AI研究实验室
166	Hermes 4 70BNous Research	10	Nous Research
167	Gemma 4 E2BDeepMind	10	DeepMind
168	Falcon-H1R-7BTII UAE	10	TII UAE
169	Qwen3-Omni-30B-A3B阿里巴巴	10	阿里巴巴
170	Step3 VL 10BStepFun	9	StepFun
171	Llama3.3-70B-InstructFacebook AI研究实验室	9	Facebook AI研究实验室
172	Ministral 3 8BMistralAI	9	MistralAI
173	Granite 4.1 30BIBM	9	IBM
174	Llama Nemotron UltraNVIDIA	9	NVIDIA
175	ERNIE-4.5-300B-A47B百度	9	百度
176	Hermes 4 405BNous Research	9	Nous Research
177	NVIDIA Nemotron 3 Nano 4BNVIDIA	9	NVIDIA
178	NVIDIA Nemotron Nano 12B v2 VLNVIDIA	9	NVIDIA
179	Gemma 4 E4BDeepMind	9	DeepMind
180	NVIDIA Nemotron Nano 9B V2NVIDIA	9	NVIDIA
181	Hermes 4 405BNous Research	9	Nous Research
182	Llama Nemotron Super 49B v1.5Meta	9	Meta
183	K2-V2 (low)MBZUAI	8	MBZUAI
184	Kimi Linear 48B A3B InstructKimi	8	Kimi
185	Llama3.1-405BFacebook AI研究实验室	8	Facebook AI研究实验室
186	LFM2.5-8B-A1BLiquid AI	8	Liquid AI
187	Ring-flash-2.0InclusionAI	8	InclusionAI
188	Olmo 3.1 32B ThinkAI2	8	AI2
189	Qwen3.5 2BAlibaba	7	Alibaba
190	C4AI Command A (202503)CohereAI	7	CohereAI
191	Llama 3.1 Nemotron 70BNVIDIA	7	NVIDIA
192	NVIDIA Nemotron 3 NanoNVIDIA	7	NVIDIA
193	NVIDIA Nemotron Nano 9B V2NVIDIA	7	NVIDIA
194	Ministral 3 3BMistral	7	Mistral
195	Hermes 4 70BNous Research	7	Nous Research
196	Granite 4.1 8BIBM	6	IBM
197	Sarvam 30B (high)Sarvam	6	Sarvam
198	Phi-4-mini-instruct (3.8B)Microsoft Azure	6	Microsoft Azure
199	Olmo 3.1 32B InstructAI2	6	AI2
200	Gemma 4 E2BDeepMind	6	DeepMind
201	R1 1776Perplexity	6	Perplexity
202	Llama 3.2-Vision-90BFacebook AI研究实验室	6	Facebook AI研究实验室
203	EXAONE 4.0 32BLG AI Research	6	LG AI Research
204	Qwen3.5 2BAlibaba	6	Alibaba
205	Qwen3.5 0.8BAlibaba	5	Alibaba
206	DeepHermes 3 - Mistral 24BNous Research	5	Nous Research
207	Jamba 1.7 LargeAI21 Labs	5	AI21 Labs
208	Granite 4.0 H SmallIBM	5	IBM
209	Qwen3-Omni-30B-A3B阿里巴巴	5	阿里巴巴
210	Granite 4.1 3BIBM	5	IBM
211	LFM2 24B A2BLiquid AI	5	Liquid AI
212	Phi-4-reasoningMicrosoft Azure	5	Microsoft Azure
213	Amazon Nova Micro亚马逊	4	亚马逊
214	MiniCPM-V 4.6 1.3BOpenBMB	4	OpenBMB
215	NVIDIA Nemotron Nano 12B v2 VLNVIDIA	4	NVIDIA
216	Phi-4-multimodal-instruct Microsoft Azure	4	Microsoft Azure
217	Jamba Reasoning 3BAI21 Labs	4	AI21 Labs
218	Gemini 3.0 FlashGoogle Deep Mind	4	Google Deep Mind
219	Olmo 3 7B ThinkAI2	4	AI2
220	Qwen3.5 0.8BAlibaba	3	Alibaba
221	Molmo 7B-DAllen Institute for AI	3	Allen Institute for AI
222	Llama 3.2-Vision-11BFacebook AI研究实验室	3	Facebook AI研究实验室
223	Exaone 4.0 1.2BLG AI Research	3	LG AI Research
224	Olmo 3 7BAI2	2	AI2
225	Exaone 4.0 1.2BLG AI Research	2	LG AI Research
226	LFM2.5-1.2B-ThinkingLiquid AI	2	Liquid AI
227	Jamba 1.7 MiniAI21 Labs	2	AI21 Labs
228	LFM2 2.6BLiquid AI	2	Liquid AI
229	LFM2.5-1.2B-InstructLiquid AI	2	Liquid AI
230	Granite 4.0 H 1BIBM	2	IBM
231	Gemma 3-270MGoogle Deep Mind	2	Google Deep Mind
232	Apertus 70B InstructSwiss AI	2	Swiss AI
233	Granite 4.0 MicroIBM	2	IBM
234	DeepHermes 3 - Llama-3.1 8BNous Research	2	Nous Research
235	Granite 4.0 1BIBM	2	IBM
236	Molmo2-8BAI2	2	AI2
237	LFM2 8B A1BLiquid AI	1	Liquid AI
238	LFM2.5-VL-1.6BLiquid AI	1	Liquid AI
239	Granite 4.0 350MIBM	1	IBM
240	Tiny Aya GlobalCohere	1	Cohere
241	Apertus 8B InstructSwiss AI	1	Swiss AI
242	Granite 4.0 H 350MIBM	1	IBM

数据仅供参考，以官方来源为准。模型名称旁的链接可跳转到 DataLearner 模型详情页。

评测基准组成（Intelligence Index v4.0）

Intelligence Index 综合10项严格的评测基准，全面衡量AI模型能力，避免单一维度的过拟合。

GDPval-AA

智能体真实任务

τ²-Bench

智能体工具调用

Terminal-Bench

智能体编程

SciCode

编程能力

AA-LCR

长上下文推理

AA-Omniscience

知识与幻觉检测

IFBench

指令遵循

Humanity's Last Exam

推理与知识

GPQA Diamond

科学推理

CritPt

物理推理

常见问题 (FAQ)

什么是 Artificial Analysis Intelligence Index？▼

Artificial Analysis Intelligence Index v4.0 是一个综合评测指数，聚合了10项具有挑战性的评估——涵盖数学、科学、编程、智能体任务和推理——以全面衡量AI能力。它旨在防止单一维度的过拟合，提供一个统一分数来追踪模型进步。

智能指数是如何计算的？▼

该指数综合了10项评测的分数：GDPval-AA（智能体真实任务）、τ²-Bench（工具调用）、Terminal-Bench Hard（智能体编程）、SciCode（编程）、AA-LCR（长上下文推理）、AA-Omniscience（知识与幻觉检测）、IFBench（指令遵循）、Humanity's Last Exam（推理）、GPQA Diamond（科学推理）和 CritPt（物理推理）。所有测试由 Artificial Analysis 在标准化硬件上独立运行。

这与 LMArena 排行榜有什么区别？▼

LMArena 排名基于众包用户投票（盲测A/B对比的Elo评分），反映主观的人类偏好。而 Artificial Analysis Intelligence Index 使用标准化的自动评测基准进行客观评分，衡量特定领域的技术能力。两者各有价值——LMArena 捕捉真实用户体验，而 AA Intelligence Index 提供可复现的技术测量。

在哪里可以找到原始数据？▼

原始排行榜和详细方法论可在 artificialanalysis.ai 查看。Intelligence Index 的方法论详见 Intelligence Index 页面。