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 Fable 5

最高得分

模型数量

261

数据版本

2026年07月12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 Fable 5Anthropic	60	Anthropic
	GPT-5.6 Sol (max)OpenAI	59	OpenAI
	GPT-5.6 Sol (xhigh)OpenAI	58	OpenAI
4	GPT-5.6 Sol (high)OpenAI	56	OpenAI
5	Claude Opus 4.8 (max)Anthropic	56	Anthropic
6	GPT-5.6 Terra (max)OpenAI	55	OpenAI
7	GPT-5.5 (xhigh)OpenAI	55	OpenAI
8	Grok 4.5 (high)SpaceXAI	54	SpaceXAI
9	GPT-5.6 Sol (medium)OpenAI	54	OpenAI
10	Opus 4.7 (max)Anthropic	54	Anthropic
11	Claude Sonnet 5 (max)Anthropic	53	Anthropic
12	GPT-5.5 (high)OpenAI	53	OpenAI
13	GPT-5.6 Terra (xhigh)OpenAI	52	OpenAI
14	GPT-5.6 Luna (max)OpenAI	51	OpenAI
15	GLM-5.2 (max)智谱AI	51	智谱AI
16	Muse Spark 1.1 (xhigh)Meta	51	Meta
17	GPT-5.5 (medium)OpenAI	50	OpenAI
18	Gemini 3.5 FlashGoogle Deep Mind	50	Google Deep Mind
19	GPT-5.6 Sol (low)OpenAI	49	OpenAI
20	GPT-5.6 Luna (xhigh)OpenAI	49	OpenAI
21	GPT-5.6 Terra (high)OpenAI	49	OpenAI
22	Gemini 3.1 Pro PreviewGoogle Deep Mind	46	Google Deep Mind
23	GPT-5.6 Luna (high)OpenAI	46	OpenAI
24	Qwen3.7 Max阿里巴巴	46	阿里巴巴
25	GPT-5.6 Terra (medium)OpenAI	46	OpenAI
26	Gemini 3.5 Flash (medium)Google	45	Google
27	MiniMax-M3MiniMax	44	MiniMax
28	DeepSeek-V4-Pro (max)DeepSeek-AI	44	DeepSeek-AI
29	GPT-5.3 Codex (xhigh)OpenAI	44	OpenAI
30	Kimi K2.6Moonshot AI	44	Moonshot AI
31	GPT-5.5 (low)OpenAI	43	OpenAI
32	Muse SparkFacebook AI研究实验室	43	Facebook AI研究实验室
33	Opus 4.7 (high)Anthropic	43	Anthropic
34	MiMo-V2.5-ProXiaomi	42	Xiaomi
35	Kimi K2.7 CodeKimi	42	Kimi
36	Claude Sonnet 5 (Non-reasoning)Anthropic	42	Anthropic
37	GPT-5.6 Sol (Non-reasoning)OpenAI	41	OpenAI
38	Nex-N2-ProNex AGI	41	Nex AGI
39	DeepSeek-V4-Pro (high)DeepSeek-AI	41	DeepSeek-AI
40	GPT-5.6 Terra (low)OpenAI	40	OpenAI
41	DeepSeek-V4-Flash (max)DeepSeek-AI	40	DeepSeek-AI
42	GLM 5.1智谱AI	40	智谱AI
43	GPT-5.4 mini (xhigh)OpenAI	40	OpenAI
44	Grok Build 0.1 0616xAI	40	xAI
45	Qwen 3.6 Plus Preview阿里巴巴	40	阿里巴巴
46	Qwen3.7 PlusAlibaba	39	Alibaba
47	JT-4.1 Flash 236B A21BChina Mobile	39	China Mobile
48	GPT-5.4 nano (xhigh)OpenAI	38	OpenAI
49	MiniMax-M2.7MiniMaxAI	38	MiniMaxAI
50	GLM-5-Turbo智谱AI	38	智谱AI
51	GPT-5.6 Luna (medium)OpenAI	38	OpenAI
52	Nemotron 3 UltraNVIDIA	38	NVIDIA
53	Grok 4.3 Beta (high)xAI	38	xAI
54	DeepSeek-V4-Flash (high)DeepSeek-AI	37	DeepSeek-AI
55	MiMo-V2.5Xiaomi	37	Xiaomi
56	Qwen3.6-27B阿里巴巴	37	阿里巴巴
57	Nova 2 Omni（Preview）亚马逊	36	亚马逊
58	Grok 4.3 Beta (medium)xAI	36	xAI
59	Grok 4.3 Beta (low)xAI	35	xAI
60	KAT-Coder-Pro V2KwaiKAT	35	KwaiKAT
61	GPT-5.5 (non-reasoning)OpenAI	35	OpenAI
62	GLM 5.1智谱AI	35	智谱AI
63	MiMo-V2-OmniXiaomi	35	Xiaomi
64	Gemini 3.5 Flash (minimal)Google Deep Mind	35	Google Deep Mind
65	KAT-Coder-Pro V1KwaiKAT	35	KwaiKAT
66	Kimi K2.6Moonshot AI	35	Moonshot AI
67	GLM-5V-Turbo智谱AI	34	智谱AI
68	Claude Sonnet 4.6 (Non-reasoning, Low Effort)Anthropic	34	Anthropic
69	GLM-5.2智谱AI	34	智谱AI
70	GPT-5.6 Terra (Non-reasoning)OpenAI	34	OpenAI
71	Qwen3.5-397B-A17B阿里巴巴	34	阿里巴巴
72	Hy3 Pre腾讯AI实验室	34	腾讯AI实验室
73	GPT-5.6 Luna (low)OpenAI	33	OpenAI
74	Gemini 2.0 Flash ExperimentalDeepMind	33	DeepMind
75	Qwen3.5-122B-A10B阿里巴巴	32	阿里巴巴
76	Qwen3.5-397B-A17B阿里巴巴	32	阿里巴巴
77	Qwen3.6-35B-A3B阿里巴巴	32	阿里巴巴
78	DeepSeek-V4-ProDeepSeek-AI	31	DeepSeek-AI
79	Qwen3.5-Omni-Plus阿里巴巴	31	阿里巴巴
80	Ring-2.6-1TInclusionAI	31	InclusionAI
81	Qwen3.6-27B阿里巴巴	30	阿里巴巴
82	OpenAI o3OpenAI	30	OpenAI
83	Step 3.7 FlashStepFun	30	StepFun
84	GPT-5.4 nanoOpenAI	30	OpenAI
85	Mistral Medium 3.5MistralAI	30	MistralAI
86	GPT-5.4 mini (medium)OpenAI	30	OpenAI
87	Haiku 4.5Anthropic	30	Anthropic
88	Gemma 4 31BDeepMind	29	DeepMind
89	GPT-5.5 Instant (June 2026)OpenAI	29	OpenAI
90	DeepSeek-V4-FlashDeepSeek-AI	29	DeepSeek-AI
91	JT-35B-FlashChina Mobile	28	China Mobile
92	MiMo-V2.5-ProXiaomi	28	Xiaomi
93	Qwen3.5-122B-A10B阿里巴巴	28	阿里巴巴
94	GPT-5.6 Luna (Non-reasoning)OpenAI	27	OpenAI
95	Hy3 Pre腾讯AI实验室	26	腾讯AI实验室
96	Ling-2.6-1TInclusionAI	26	InclusionAI
97	Step 3.5 FlashStepFunAI	26	StepFunAI
98	Doubao Seed CodeByteDance Seed	26	ByteDance Seed
99	Gemini 2.5-ProGoogle Deep Mind	26	Google Deep Mind
100	Gemma 4 26B A4BDeepMind	26	DeepMind
101	NVIDIA Nemotron 3 SuperNVIDIA	25	NVIDIA
102	Mercury 2Inception	25	Inception
103	Gemini 3.1 Flash-LiteGoogle	25	Google
104	Grok 4.3 (Non-reasoning)xAI	25	xAI
105	K-EXAONELG AI Research	25	LG AI Research
106	MiMo-V2-FlashXiaomi	25	Xiaomi
107	Trinity Large ThinkingArcee AI	24	Arcee AI
108	Qwen3.6-35B-A3B阿里巴巴	24	阿里巴巴
109	GPT OSS 120B (high)OpenAI	24	OpenAI
110	Haiku 4.5Anthropic	24	Anthropic
111	Qwen3.5-35B-A3B阿里巴巴	23	阿里巴巴
112	EXAONE 4.5 33BLG AI Research	23	LG AI Research
113	C4AI Command A (202503)CohereAI	23	CohereAI
114	Gemma 4 12BGoogle	22	Google
115	ERNIE 5.0百度	22	百度
116	Gemma 4 31BDeepMind	22	DeepMind
117	Nova 2 Pro（Preview） (medium)亚马逊	22	亚马逊
118	Qwen3.5-9B-Instruct阿里巴巴	21	阿里巴巴
119	Nemotron Cascade 2 30B A3BNVIDIA	21	NVIDIA
120	Qwen3-Coder-Next阿里巴巴	21	阿里巴巴
121	Nova 2 Omni（Preview） (medium)亚马逊	21	亚马逊
122	North Mini CodeCohere	21	Cohere
123	Apriel-v1.6-15B-ThinkerServiceNow	21	ServiceNow
124	Qwen3.5-9B-Instruct阿里巴巴	20	阿里巴巴
125	Gemma 4 26B A4BDeepMind	20	DeepMind
126	Qwen3.5 4BAlibaba	20	Alibaba
127	Nova 2 Pro（Preview） (low)亚马逊	20	亚马逊
128	Mistral Small 4Mistral	20	Mistral
129	Devstral 2Mistral	19	Mistral
130	Nova 2 Lite (medium)亚马逊	19	亚马逊
131	Qwen3.5-Omni-Flash阿里巴巴	19	阿里巴巴
132	JT-MINIChina Mobile	19	China Mobile
133	Nova 2 Lite (high)亚马逊	18	亚马逊
134	Magistral Medium 1.2Mistral	18	Mistral
135	Nova 2 Lite (low)亚马逊	18	亚马逊
136	HyperNova 60B 2605Multiverse Computing	18	Multiverse Computing
137	GPT OSS 120B (low)OpenAI	18	OpenAI
138	GPT-5.4 nanoOpenAI	18	OpenAI
139	Devstral Small 2Mistral	17	Mistral
140	K2 Think V2MBZUAI	17	MBZUAI
141	LongCat Flash LiteLongCat	17	LongCat
142	HyperCLOVA X SEED Think (32B)Naver	17	Naver
143	K-EXAONELG AI Research	17	LG AI Research
144	Qwen3-Next阿里巴巴	17	阿里巴巴
145	GPT-5.4 miniOpenAI	17	OpenAI
146	Nova 2 Omni（Preview） (low)亚马逊	17	亚马逊
147	Mi:dm K 2.5 ProKorea Telecom	16	Korea Telecom
148	Qwen3.5 4BAlibaba	16	Alibaba
149	Mistral Large 3MistralAI	16	MistralAI
150	INTELLECT-3Prime Intellect	16	Prime Intellect
151	Solar Open 100BUpstage	15	Upstage
152	Qwen3-Omni-30B-A3B (reasoning)阿里巴巴	15	阿里巴巴
153	GPT OSS 20B (high)OpenAI	15	OpenAI
154	Nova 2 Pro（Preview）亚马逊	14	亚马逊
155	GPT OSS 20B (low)OpenAI	14	OpenAI
156	Llama 4 MaverickFacebook AI研究实验室	14	Facebook AI研究实验室
157	K2-V2 (high)MBZUAI	14	MBZUAI
158	NVIDIA Nemotron 3 NanoNVIDIA	14	NVIDIA
159	Solar Pro 3Upstage	14	Upstage
160	Ling 2.6 FlashInclusionAI	14	InclusionAI
161	Qwen3-Next阿里巴巴	14	阿里巴巴
162	Tri-21B-think PreviewTrillion Labs	14	Trillion Labs
163	DiffusionGemma 26B A4BGoogle	13	Google
164	Gemma 4 12B (Non-reasoning)Google	13	Google
165	Motif-2-12.7BMotif Technologies	13	Motif Technologies
166	Nova PremierAmazon	13	Amazon
167	Gemma 4 E4BDeepMind	12	DeepMind
168	K2-V2 (medium)MBZUAI	12	MBZUAI
169	Llama Nemotron Super 49B v1.5Meta	12	Meta
170	Mistral Small 4Mistral	12	Mistral
171	Tri-21B-ThinkTrillion Labs	12	Trillion Labs
172	MiniCPM5-1BOpenBMB	12	OpenBMB
173	Sarvam 105B (high)Sarvam	12	Sarvam
174	Nova 2 Lite亚马逊	12	亚马逊
175	MiniCPM5-1BOpenBMB	12	OpenBMB
176	Magistral Small 1.2Mistral	11	Mistral
177	Nanbeige4.1-3BNanbeige	11	Nanbeige
178	Ministral 3 14BMistralAI	11	MistralAI
179	EXAONE 4.0 32BLG AI Research	11	LG AI Research
180	Nova 2 Omni（Preview）亚马逊	11	亚马逊
181	Llama 4 ScoutFacebook AI研究实验室	10	Facebook AI研究实验室
182	Hermes 4 70BNous Research	10	Nous Research
183	Falcon-H1R-7BTII UAE	10	TII UAE
184	Qwen3-Omni-30B-A3B阿里巴巴	10	阿里巴巴
185	Step3 VL 10BStepFun	9	StepFun
186	Llama3.3-70B-InstructFacebook AI研究实验室	9	Facebook AI研究实验室
187	Gemma 4 E2BDeepMind	9	DeepMind
188	Llama Nemotron UltraNVIDIA	9	NVIDIA
189	ERNIE-4.5-300B-A47B百度	9	百度
190	Hermes 4 405BNous Research	9	Nous Research
191	Solar Pro 2Upstage	9	Upstage
192	NVIDIA Nemotron Nano 12B v2 VLNVIDIA	9	NVIDIA
193	Ministral 3 8BMistralAI	9	MistralAI
194	Gemma 4 E4BDeepMind	9	DeepMind
195	Granite 4.1 30BIBM	9	IBM
196	NVIDIA Nemotron Nano 9B V2NVIDIA	9	NVIDIA
197	Hermes 4 405BNous Research	9	Nous Research
198	NVIDIA Nemotron 3 Nano 4BNVIDIA	9	NVIDIA
199	Llama Nemotron Super 49B v1.5Meta	9	Meta
200	K2-V2 (low)MBZUAI	9	MBZUAI
201	Kimi Linear 48B A3B InstructKimi	9	Kimi
202	Llama3.1-405BFacebook AI研究实验室	9	Facebook AI研究实验室
203	LFM2.5-8B-A1BLiquid AI	8	Liquid AI
204	Ring-flash-2.0InclusionAI	8	InclusionAI
205	Olmo 3.1 32B ThinkAI2	8	AI2
206	Solar Pro 2Upstage	8	Upstage
207	C4AI Command A (202503)CohereAI	8	CohereAI
208	Qwen3.5 2BAlibaba	8	Alibaba
209	Llama 3.1 Nemotron 70BNVIDIA	8	NVIDIA
210	NVIDIA Nemotron 3 NanoNVIDIA	7	NVIDIA
211	NVIDIA Nemotron Nano 9B V2NVIDIA	7	NVIDIA
212	Hermes 4 70BNous Research	7	Nous Research
213	Ministral 3 3BMistral	7	Mistral
214	Granite 4.1 8BIBM	7	IBM
215	Sarvam 30B (high)Sarvam	7	Sarvam
216	Olmo 3.1 32B InstructAI2	6	AI2
217	Gemma 4 E2BDeepMind	6	DeepMind
218	R1 1776Perplexity	6	Perplexity
219	Llama 3.2-Vision-90BFacebook AI研究实验室	6	Facebook AI研究实验室
220	Phi-4-mini-instruct (3.8B)Microsoft Azure	6	Microsoft Azure
221	EXAONE 4.0 32BLG AI Research	6	LG AI Research
222	Qwen3.5 2BAlibaba	6	Alibaba
223	Qwen3.5 0.8BAlibaba	5	Alibaba
224	DeepHermes 3 - Mistral 24BNous Research	5	Nous Research
225	Jamba 1.7 LargeAI21 Labs	5	AI21 Labs
226	Granite 4.0 H SmallIBM	5	IBM
227	Qwen3-Omni-30B-A3B阿里巴巴	5	阿里巴巴
228	LFM2 24B A2BLiquid AI	5	Liquid AI
229	Phi 4 - 14BMicrosoft Azure	5	Microsoft Azure
230	Amazon Nova Micro亚马逊	5	亚马逊
231	Granite 4.1 3BIBM	5	IBM
232	NVIDIA Nemotron Nano 12B v2 VLNVIDIA	5	NVIDIA
233	Phi-4-multimodal-instruct Microsoft Azure	5	Microsoft Azure
234	Qwen3.5 0.8BAlibaba	4	Alibaba
235	MiniCPM-V 4.6 1.3BOpenBMB	4	OpenBMB
236	Jamba Reasoning 3BAI21 Labs	4	AI21 Labs
237	Gemini 3.0 FlashGoogle Deep Mind	4	Google Deep Mind
238	Olmo 3 7B ThinkAI2	4	AI2
239	Molmo 7B-DAllen Institute for AI	4	Allen Institute for AI
240	Ling-mini-2.0InclusionAI	4	InclusionAI
241	Llama 3.2-Vision-11BFacebook AI研究实验室	3	Facebook AI研究实验室
242	Exaone 4.0 1.2BLG AI Research	3	LG AI Research
243	Olmo 3 7BAI2	3	AI2
244	Exaone 4.0 1.2BLG AI Research	3	LG AI Research
245	LFM2.5-1.2B-ThinkingLiquid AI	3	Liquid AI
246	Jamba 1.7 MiniAI21 Labs	3	AI21 Labs
247	LFM2 2.6BLiquid AI	3	Liquid AI
248	LFM2.5-1.2B-InstructLiquid AI	3	Liquid AI
249	Granite 4.0 H 1BIBM	3	IBM
250	Gemma 3-270MGoogle Deep Mind	2	Google Deep Mind
251	Apertus 70B InstructSwiss AI	2	Swiss AI
252	Granite 4.0 MicroIBM	2	IBM
253	DeepHermes 3 - Llama-3.1 8BNous Research	2	Nous Research
254	Granite 4.0 1BIBM	2	IBM
255	Molmo2-8BAI2	2	AI2
256	LFM2 8B A1BLiquid AI	2	Liquid AI
257	LFM2.5-VL-1.6BLiquid AI	1	Liquid AI
258	Granite 4.0 350MIBM	1	IBM
259	Tiny Aya GlobalCohere	1	Cohere
260	Apertus 8B InstructSwiss AI	1	Swiss AI
261	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。