Artificial Analysis Intelligence Index leaderboard

A display-only intelligence index published by Artificial Analysis that aggregates provider-reported and benchmark-derived signals into a single model-level score.

Models ranked by their published Artificial Analysis Intelligence Index value

Rank	Model	Provider	Aggregated model score
1	GPT-5.6 Sol	OpenAI	58.9
2	GPT-5.6 Terra	OpenAI	55.0
3	Claude Fable 5.1	Anthropic	53.4
4	DeepSeek V4 Pro 0813	DeepSeek	53.2
5	GPT-6 Astra	OpenAI	52.8
6	GPT-5.6 Luna	OpenAI	51.2
7	Claude Opus 5	Anthropic	50.7
8	Gemini 3.5 Flash	Google	50.2
9	Claude Fable 5	Anthropic	49.7
10	Muse Spark 1.3	Meta	48.2
11	Qwen3.8 Max Preview	Alibaba	45.4
12	GLM-5.3	Z.AI	44.9
13	Grok 4.6	xAI	44.4
14	Kimi K3	Moonshot AI	43.8
15	Claude Opus 4.8	Anthropic	42.0
16	GLM-5.3-Flash	Z.AI	41.9
17	Hy3 Preview	Tencent	41.2
18	Gemini 3.8 Flash	Google	41.2
19	Claude Opus 4.7 (Adaptive)	Anthropic	40.7
20	Qwen3.8-Flash-Next	Alibaba	39.9
21	Muse Spark 1.2	Meta	39.8
22	DeepSeek V4.1 Flash	DeepSeek	39.5
23	Gemini 3.7 Flash	Google	39.4
24	Grok 4.5	xAI	39.1
25	GPT-5.4	OpenAI	39.0
26	GPT-5.5	OpenAI	38.6
27	Claude Sonnet 5	Anthropic	38.4
28	Grok 4.3	xAI	37.6
29	DeepSeek V4 Flash 0731	DeepSeek	34.5
30	Gemini 3.6 Flash	Google	34.3
31	Muse Spark 1.1	Meta	34.3
32	GLM-5.2	Z.AI	34.0
33	Qwen3.8-27B	Alibaba	33.9
34	GPT-5.3 Codex	OpenAI	32.5
35	Claude Opus 4.6 (Adaptive)	Anthropic	31.9
36	Muse Spark	Meta	31.3
37	Claude Opus 4.7	Anthropic	30.9
38	GPT-5.2	OpenAI	30.4
39	APApodex 1.1	Apodex	30.4
39	APApodex 1.1 Mini	Apodex	30.4
41	Gemini 3.1 Pro	Google	30.4
42	Qwen3.7 Max	Alibaba	29.9
43	MiniMax M3	MiniMax	29.6
44	Claude Opus 4.5 Thinking	Anthropic	29.1
45	MiMo-V2-Pro	Xiaomi	28.6
46	GPT-5.2-Codex	OpenAI	28.5
47	Qwen 3.6 Max (preview)	Alibaba	28.4
48	Solar Pro 4	Upstage	28.1
49	Gemini 3 Pro	Google	28.0
50	GLM-5	Z.AI	27.9
51	Kimi K2.6	Moonshot AI	27.5
52	MCQuasar 438B	Multiverse Computing	27.1
53	Qwen3.6 Plus	Alibaba	27.0
54	GLM-5-Turbo	Z.AI	26.6
55	GLM-5.1	Z.AI	26.4
56	MiMo-V2.5-Pro	Xiaomi	26.4
57	Claude Opus 4.6	Anthropic	26.4
58	Kimi K2.7 Code	Moonshot AI	26.3
59	TMInkling-Small	Thinking Machines Lab	26.1
60	Qwen3.7 Plus	Alibaba	25.8
61	Hy3	Tencent	25.8
62	TMInkling	Thinking Machines Lab	25.5
63	Ling 3.0 Flash VL	InclusionAI	25.0
64	GPT-5.1	OpenAI	24.7
65	Claude Sonnet 4.6	Anthropic	24.7
66	GPT-5.4 mini	OpenAI	24.6
67	MiMo-V2-Omni	Xiaomi	23.9
68	GPT-5.1-Codex	OpenAI	23.7
68	GPT-5.1-Codex-Max	OpenAI	23.7
70	Claude Opus 4.5	Anthropic	23.7
71	GLM-5V-Turbo	Z.AI	23.5
72	Kimi K2.5	Moonshot AI	23.5
72	Kimi K2.5 (Reasoning)	Moonshot AI	23.5
74	Nemotron 3 Ultra	NVIDIA	23.4
75	MiniMax M2.7	MiniMax	23.2
76	GPT-5 (high)	OpenAI	23.0
77	Qwen3.5-27B	Alibaba	22.9
78	GPT-5 (medium)	OpenAI	22.9
79	Claude 4.1 Opus Thinking	Anthropic	22.9
80	STA.X K2	SK Telecom	22.7
81	Gemini 3.5 Flash-Lite	Google	22.7
82	Command A+	Cohere	22.5
83	Grok 4	xAI	22.5
84	GLM-4.7	Z.AI	22.2
85	Qwen3.6-27B	Alibaba	21.9
86	o3-pro	OpenAI	21.9
87	Qwen3.5 397B	Alibaba	21.4
87	Qwen3.5 397B (Reasoning)	Alibaba	21.4
89	GPT-5.4 nano	OpenAI	21.2
90	Ling 3.0 Flash	InclusionAI	20.6
90	Ling 3.0 Flash FP8	InclusionAI	20.6
92	Grok 4.1 Fast (Reasoning)	xAI	20.4
93	o3	OpenAI	20.2
94	K-EXAONE 2.0	LG AI Research	19.7
95	Step 3.7 Flash	StepFun	19.5
96	Qwen3.5-35B-A3B	Alibaba	19.3
97	Qwen3.6-35B-A3B	Alibaba	18.8
98	Claude 4.1 Opus	Anthropic	18.6
99	Muse Glimmer 30B	Meta	18.1
100	Grok 4 Fast (Reasoning)	xAI	17.9
101	Gemini 3 Flash	Google	17.9
102	Gemma 4 26B A4B	Google	16.7
103	Gemini 2.5 Pro	Google	16.7
104	Claude 4 Sonnet	Anthropic	16.6
105	Qwen3.5-122B-A10B	Alibaba	16.2
106	MiMo-V2-Flash	Xiaomi	16.1
107	DeepSeek V3.2	DeepSeek	16.0
108	Qwen3 Max	Alibaba	15.6
109	Gemma 4 31B	Google	15.4
110	o1	OpenAI	15.2
111	GLM-4.6	Z.AI	14.9
112	Mistral Medium 3.5 128B	Mistral	14.9
113	Granite 4.2 30B	IBM	14.8
114	K-Exaone	LG AI Research	14.4
115	Gemma 4 12B	Google	14.2
116	Grok Code Fast 1	xAI	14.1
117	Ling 2.6 Flash	InclusionAI	14.1
118	DeepSeek V3.1	DeepSeek	13.7
119	Nemotron 3.5 Lightning 30B A3B NVFP4	NVIDIA	13.6
120	Nemotron 3 Super 100B	NVIDIA	13.6
121	DeepSeek V3.1 (Reasoning)	DeepSeek	13.5
122	OPMiniCPM5-2B	OpenBMB	13.1
123	DeepSeek-R1	DeepSeek	13.1
124	Kimi K2	Moonshot AI	12.7
125	GPT-4.1	OpenAI	12.7
126	o3-mini	OpenAI	12.5
127	o1-pro	OpenAI	12.4
128	GPT-OSS 120B	OpenAI	12.3
129	Ling 3.0 Tiny	InclusionAI	11.9
130	Granite 4.2 8B	IBM	11.8
131	Mistral Small 4	Mistral	11.4
131	Mistral Small 4 (Reasoning)	Mistral	11.4
133	o1-preview	OpenAI	11.4
134	Grok 4.1 Fast	xAI	11.3
135	GLM-4.5-Air	Z.AI	11.1
136	Trinity-Large-Preview	Arcee AI	10.9
136	Trinity-Large-Thinking	Arcee AI	10.9
138	Nemotron 3 Nano Omni 30B A3B	NVIDIA	10.3
139	GPT-4.1 mini	OpenAI	10.2
140	North Mini Code	Cohere	9.9
141	Gemini 2.5 Flash	Google	9.8
142	Mistral Large 3	Mistral	9.7
143	Llama 4 Maverick	Meta	9.3
144	Granite 4.2 3B	IBM	9.1
145	Mistral Medium 3	Mistral	9.1
146	GPT-OSS 20B	OpenAI	9.0
147	Gemma 4 E4B	Google	8.9
148	Nemotron 3 Nano 30B	NVIDIA	8.9
149	SASarvam 105B	Sarvam	8.8
150	Claude 3 Opus	Anthropic	8.7
151	DeepSeek V3	DeepSeek	8.5
152	GPT-4o	OpenAI	8.4
153	LFM2.5-2.6B	LiquidAI	8.4
154	DeepSeek R1 Distill Qwen 32B	DeepSeek	8.4
155	Gemini 1.5 Pro	Google	7.9
156	GPT-4.1 nano	OpenAI	7.8
156	Solar Pro 3	Upstage	7.8
158	Gemma 4 E2B	Google	7.8
159	Qwen3-Omni-30B-A3B-Thinking	Alibaba	7.8
160	FAUltravox v0.6 Llama 3.3 70B	Fixie AI	7.7
161	Mistral Large 2	Mistral	7.6
162	Nemotron Ultra 253B	NVIDIA	7.5
163	Llama 3.1 405B	Meta	7.3
164	LFM2.5-8B-A1B	LiquidAI	7.2
165	GPT-4 Turbo	OpenAI	7.0
166	Solar Pro 2	Upstage	7.0
167	Nova Pro	Amazon	7.0
168	Qwen2.5 Coder 32B Instruct	Alibaba	6.7
169	GPT-4o mini	OpenAI	6.7
170	SASarvam 30B	Sarvam	6.6
171	Llama 4 Scout	Meta	6.5
172	CECeleris-1	Celeris	6.3
173	Exaone 4.0 32B	LG AI Research	6.3
174	Qwen3-Omni-30B-A3B-Instruct	Alibaba	6.0
175	Phi-4	Microsoft	5.9
176	Phi-4 Multimodal Instruct	Microsoft	5.8
177	Claude 3 Haiku	Anthropic	5.6
178	Gemini 1.0 Pro	Google	5.3
179	Exaone 4.0 1.2B	LG AI Research	5.2
180	Granite-4.0-H-1B	IBM	5.2
181	Granite-4.0-1B	IBM	5.0
182	Gemma 3 27B	Google	4.8
183	Granite-4.0-350M	IBM	4.8
183	Granite-4.0-H-350M	IBM	4.8
183	LFM2.5-VL-1.6B-Extract	LiquidAI	4.8

Evidence key: ObservedLast good

Rows are ordered by the value Artificial Analysis published, highest first. The source does not state whether a higher value is the better result, so this page does not either: for a benchmark that measures a rate of failure, read the table from the bottom.