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

214

数据版本

2026年06月19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
	Claude Opus 4.8 (max)Anthropic	56	Anthropic
	GPT-5.5 (xhigh)OpenAI	55	OpenAI
4	Opus 4.7 (max)Anthropic	54	Anthropic
5	GPT-5.5 (high)OpenAI	53	OpenAI
6	GLM-5.2 (max)智谱AI	51	智谱AI
7	Gemini 3.5 FlashGoogle Deep Mind	50	Google Deep Mind
8	Claude Sonnet 4.6 (max)Anthropic	47	Anthropic
9	GPT-5.5 (medium)OpenAI	47	OpenAI
10	Gemini 3.1 Pro PreviewGoogle Deep Mind	46	Google Deep Mind
11	Qwen3.7 Max阿里巴巴	46	阿里巴巴
12	Gemini 3.5 Flash (medium)Google	45	Google
13	MiniMax-M3MiniMax	44	MiniMax
14	DeepSeek-V4-Pro (max)DeepSeek-AI	44	DeepSeek-AI
15	GPT-5.3 Codex (xhigh)OpenAI	44	OpenAI
16	Muse SparkFacebook AI研究实验室	43	Facebook AI研究实验室
17	Kimi K2.6Moonshot AI	43	Moonshot AI
18	Opus 4.7 (high)Anthropic	43	Anthropic
19	MiMo-V2.5-ProXiaomi	42	Xiaomi
20	Kimi K2.7 CodeKimi	42	Kimi
21	GPT-5.5 (low)OpenAI	42	OpenAI
22	DeepSeek-V4-Pro (high)DeepSeek-AI	41	DeepSeek-AI
23	DeepSeek-V4-Flash (max)DeepSeek-AI	40	DeepSeek-AI
24	GLM 5.1智谱AI	40	智谱AI
25	MiMo-V2.5Xiaomi	40	Xiaomi
26	GPT-5.4 mini (xhigh)OpenAI	40	OpenAI
27	Qwen 3.6 Plus Preview阿里巴巴	40	阿里巴巴
28	Qwen3.7 PlusAlibaba	39	Alibaba
29	GPT-5.4 nano (xhigh)OpenAI	38	OpenAI
30	MiniMax-M2.7MiniMaxAI	38	MiniMaxAI
31	GLM-5-Turbo智谱AI	38	智谱AI
32	Nemotron 3 UltraNVIDIA	38	NVIDIA
33	Grok 4.3 Beta (high)xAI	38	xAI
34	DeepSeek-V4-Flash (high)DeepSeek-AI	37	DeepSeek-AI
35	Qwen3.6-27B阿里巴巴	37	阿里巴巴
36	Nova 2 Omni（Preview）亚马逊	36	亚马逊
37	Grok 4.3 Beta (medium)xAI	36	xAI
38	Claude Sonnet 4.6 (non-reasoning)Anthropic	36	Anthropic
39	Grok 4.3 Beta (low)xAI	35	xAI
40	GLM 5.1智谱AI	35	智谱AI
41	MiMo-V2-OmniXiaomi	35	Xiaomi
42	Gemini 3.5 Flash (minimal)Google Deep Mind	35	Google Deep Mind
43	Kimi K2.6Moonshot AI	35	Moonshot AI
44	GLM-5V-Turbo智谱AI	34	智谱AI
45	Claude Sonnet 4.6 (Non-reasoning, Low Effort)Anthropic	34	Anthropic
46	Qwen3.5-397B-A17B阿里巴巴	34	阿里巴巴
47	Hy3 Pre腾讯AI实验室	34	腾讯AI实验室
48	GPT-5.5 Instant (May 2026)OpenAI	34	OpenAI
49	Gemini 2.0 Flash ExperimentalDeepMind	33	DeepMind
50	GPT-5.5 (non-reasoning)OpenAI	33	OpenAI
51	Qwen3.5-122B-A10B阿里巴巴	32	阿里巴巴
52	Qwen3.5-397B-A17B阿里巴巴	32	阿里巴巴
53	Qwen3.6-35B-A3B阿里巴巴	32	阿里巴巴
54	DeepSeek-V4-ProDeepSeek-AI	31	DeepSeek-AI
55	Qwen3.5-Omni-Plus阿里巴巴	31	阿里巴巴
56	Ring-2.6-1TInclusionAI	31	InclusionAI
57	OpenAI o3OpenAI	30	OpenAI
58	GPT-5.4 nanoOpenAI	30	OpenAI
59	Mistral Medium 3.5MistralAI	30	MistralAI
60	GPT-5.4 mini (medium)OpenAI	30	OpenAI
61	Step 3.7 FlashStepFun	30	StepFun
62	Haiku 4.5Anthropic	30	Anthropic
63	Gemma 4 31BDeepMind	29	DeepMind
64	C4AI Command A (202503)CohereAI	29	CohereAI
65	Qwen3.6-27B阿里巴巴	29	阿里巴巴
66	DeepSeek-V4-FlashDeepSeek-AI	29	DeepSeek-AI
67	JT-35B-FlashChina Mobile	28	China Mobile
68	Qwen3.5-122B-A10B阿里巴巴	28	阿里巴巴
69	MiMo-V2.5-ProXiaomi	28	Xiaomi
70	Gemini 2.5-ProGoogle Deep Mind	27	Google Deep Mind
71	Hy3 Pre腾讯AI实验室	26	腾讯AI实验室
72	Ling-2.6-1TInclusionAI	26	InclusionAI
73	Step 3.5 FlashStepFunAI	26	StepFunAI
74	Doubao Seed CodeByteDance Seed	26	ByteDance Seed
75	Gemma 4 26B A4BDeepMind	26	DeepMind
76	NVIDIA Nemotron 3 SuperNVIDIA	25	NVIDIA
77	Mercury 2Inception	25	Inception
78	Gemini 3.1 Flash-LiteGoogle	25	Google
79	Qwen3.5-9B-Instruct阿里巴巴	25	阿里巴巴
80	Gemma 4 31BDeepMind	25	DeepMind
81	Grok 4.3 (Non-reasoning)xAI	25	xAI
82	K-EXAONELG AI Research	25	LG AI Research
83	Trinity Large ThinkingArcee AI	24	Arcee AI
84	Qwen3.6-35B-A3B阿里巴巴	24	阿里巴巴
85	GPT OSS 120B (high)OpenAI	24	OpenAI
86	Haiku 4.5Anthropic	24	Anthropic
87	Qwen3.5-35B-A3B阿里巴巴	23	阿里巴巴
88	MiMo-V2-FlashXiaomi	23	Xiaomi
89	EXAONE 4.5 33BLG AI Research	23	LG AI Research
90	HyperNova 60B 2605Multiverse Computing	22	Multiverse Computing
91	Gemma 4 12BGoogle	22	Google
92	ERNIE 5.0百度	22	百度
93	Nova 2 Pro（Preview） (medium)亚马逊	22	亚马逊
94	Nemotron Cascade 2 30B A3BNVIDIA	21	NVIDIA
95	Qwen3-Coder-Next阿里巴巴	21	阿里巴巴
96	Nova 2 Omni（Preview） (medium)亚马逊	21	亚马逊
97	Mistral Small 4Mistral	21	Mistral
98	North Mini CodeCohere	21	Cohere
99	Nova 2 Lite (high)亚马逊	21	亚马逊
100	Qwen3.5-9B-Instruct阿里巴巴	20	阿里巴巴
101	Magistral Medium 1.2Mistral	20	Mistral
102	Gemma 4 26B A4BDeepMind	20	DeepMind
103	Qwen3.5 4BAlibaba	20	Alibaba
104	Qwen3-Next阿里巴巴	20	阿里巴巴
105	Nova 2 Pro（Preview） (low)亚马逊	20	亚马逊
106	Ling 2.6 FlashInclusionAI	19	InclusionAI
107	Nova 2 Lite (medium)亚马逊	19	亚马逊
108	Qwen3.5-Omni-Flash阿里巴巴	19	阿里巴巴
109	JT-MINIChina Mobile	19	China Mobile
110	Nova 2 Lite (low)亚马逊	18	亚马逊
111	GPT OSS 120B (low)OpenAI	18	OpenAI
112	GPT-5.4 nanoOpenAI	18	OpenAI
113	NVIDIA Nemotron 3 NanoNVIDIA	18	NVIDIA
114	LongCat Flash LiteLongCat	17	LongCat
115	K-EXAONELG AI Research	17	LG AI Research
116	GPT-5.4 miniOpenAI	17	OpenAI
117	Nova 2 Omni（Preview） (low)亚马逊	17	亚马逊
118	Nova 2 Pro（Preview）亚马逊	16	亚马逊
119	Mi:dm K 2.5 ProKorea Telecom	16	Korea Telecom
120	Mistral Large 3MistralAI	16	MistralAI
121	Qwen3.5 4BAlibaba	16	Alibaba
122	INTELLECT-3Prime Intellect	16	Prime Intellect
123	Devstral 2Mistral	15	Mistral
124	Solar Open 100BUpstage	15	Upstage
125	Qwen3-Omni-30B-A3B (reasoning)阿里巴巴	15	阿里巴巴
126	GPT OSS 20B (high)OpenAI	15	OpenAI
127	GPT OSS 20B (low)OpenAI	14	OpenAI
128	Llama 4 MaverickFacebook AI研究实验室	14	Facebook AI研究实验室
129	Solar Pro 3Upstage	14	Upstage
130	Qwen3-Next阿里巴巴	14	阿里巴巴
131	Gemma 4 12B (Non-reasoning)Google	13	Google
132	Devstral Small 2Mistral	13	Mistral
133	Motif-2-12.7BMotif Technologies	13	Motif Technologies
134	Nova PremierAmazon	13	Amazon
135	Gemma 4 E4BDeepMind	12	DeepMind
136	Llama Nemotron Super 49B v1.5Meta	12	Meta
137	Mistral Small 4Mistral	12	Mistral
138	MiniCPM5-1BOpenBMB	12	OpenBMB
139	Magistral Small 1.2Mistral	12	Mistral
140	Sarvam 105B (high)Sarvam	12	Sarvam
141	Nova 2 Lite亚马逊	12	亚马逊
142	MiniCPM5-1BOpenBMB	12	OpenBMB
143	EXAONE 4.0 32BLG AI Research	11	LG AI Research
144	Nova 2 Omni（Preview）亚马逊	11	亚马逊
145	Qwen3.5 2BAlibaba	10	Alibaba
146	Nanbeige4.1-3BNanbeige	10	Nanbeige
147	Llama 4 ScoutFacebook AI研究实验室	10	Facebook AI研究实验室
148	Ministral 3 14BMistralAI	10	MistralAI
149	Falcon-H1R-7BTII UAE	10	TII UAE
150	Qwen3-Omni-30B-A3B阿里巴巴	10	阿里巴巴
151	Step3 VL 10BStepFun	9	StepFun
152	Gemma 4 E2BDeepMind	9	DeepMind
153	Llama Nemotron UltraNVIDIA	9	NVIDIA
154	ERNIE-4.5-300B-A47B百度	9	百度
155	Solar Pro 2Upstage	9	Upstage
156	NVIDIA Nemotron Nano 12B v2 VLNVIDIA	9	NVIDIA
157	Ministral 3 8BMistralAI	9	MistralAI
158	Gemma 4 E4BDeepMind	9	DeepMind
159	Granite 4.1 30BIBM	9	IBM
160	NVIDIA Nemotron Nano 9B V2NVIDIA	9	NVIDIA
161	NVIDIA Nemotron 3 Nano 4BNVIDIA	9	NVIDIA
162	Qwen3.5 2BAlibaba	9	Alibaba
163	Llama Nemotron Super 49B v1.5Meta	9	Meta
164	Llama3.3-70B-InstructFacebook AI研究实验室	9	Facebook AI研究实验室
165	Kimi Linear 48B A3B InstructKimi	9	Kimi
166	Llama3.1-405BFacebook AI研究实验室	9	Facebook AI研究实验室
167	LFM2.5-8B-A1BLiquid AI	8	Liquid AI
168	Ring-flash-2.0InclusionAI	8	InclusionAI
169	Solar Pro 2Upstage	8	Upstage
170	C4AI Command A (202503)CohereAI	8	CohereAI
171	Llama 3.1 Nemotron 70BNVIDIA	8	NVIDIA
172	NVIDIA Nemotron 3 NanoNVIDIA	7	NVIDIA
173	NVIDIA Nemotron Nano 9B V2NVIDIA	7	NVIDIA
174	Granite 4.1 8BIBM	7	IBM
175	Sarvam 30B (high)Sarvam	7	Sarvam
176	Gemma 4 E2BDeepMind	6	DeepMind
177	R1 1776Perplexity	6	Perplexity
178	Llama 3.2-Vision-90BFacebook AI研究实验室	6	Facebook AI研究实验室
179	EXAONE 4.0 32BLG AI Research	6	LG AI Research
180	Ministral 3 3BMistral	6	Mistral
181	Jamba 1.7 LargeAI21 Labs	5	AI21 Labs
182	Granite 4.0 H SmallIBM	5	IBM
183	Qwen3-Omni-30B-A3B阿里巴巴	5	阿里巴巴
184	Qwen3.5 0.8BAlibaba	5	Alibaba
185	LFM2 24B A2BLiquid AI	5	Liquid AI
186	Phi 4 - 14BMicrosoft Azure	5	Microsoft Azure
187	Amazon Nova Micro亚马逊	5	亚马逊
188	NVIDIA Nemotron Nano 12B v2 VLNVIDIA	5	NVIDIA
189	Phi-4-multimodal-instruct Microsoft Azure	5	Microsoft Azure
190	Qwen3.5 0.8BAlibaba	4	Alibaba
191	MiniCPM-V 4.6 1.3BOpenBMB	4	OpenBMB
192	Jamba Reasoning 3BAI21 Labs	4	AI21 Labs
193	Gemini 3.0 FlashGoogle Deep Mind	4	Google Deep Mind
194	Ling-mini-2.0InclusionAI	4	InclusionAI
195	Llama 3.2-Vision-11BFacebook AI研究实验室	3	Facebook AI研究实验室
196	Granite 4.1 3BIBM	3	IBM
197	Phi-4-mini-instruct (3.8B)Microsoft Azure	3	Microsoft Azure
198	Exaone 4.0 1.2BLG AI Research	3	LG AI Research
199	Exaone 4.0 1.2BLG AI Research	3	LG AI Research
200	LFM2.5-1.2B-ThinkingLiquid AI	3	Liquid AI
201	Jamba 1.7 MiniAI21 Labs	3	AI21 Labs
202	LFM2 2.6BLiquid AI	3	Liquid AI
203	LFM2.5-1.2B-InstructLiquid AI	3	Liquid AI
204	Granite 4.0 H 1BIBM	3	IBM
205	Gemma 3-270MGoogle Deep Mind	2	Google Deep Mind
206	Apertus 70B InstructSwiss AI	2	Swiss AI
207	Granite 4.0 MicroIBM	2	IBM
208	Granite 4.0 1BIBM	2	IBM
209	LFM2 8B A1BLiquid AI	2	Liquid AI
210	LFM2.5-VL-1.6BLiquid AI	1	Liquid AI
211	Granite 4.0 350MIBM	1	IBM
212	Tiny Aya GlobalCohere	1	Cohere
213	Apertus 8B InstructSwiss AI	1	Swiss AI
214	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。