模型排行

排名	變化	模型	得分	置信區間	投票	價格 $/M
1	—	o1-preview閉源OpenAI	1,300	±4	26,268	—
2	—	claude-3-5-sonnet-20241022閉源Anthropic · 另有 1 個檔位	1,284	±5	26,047	—
3	—	gemini-1.5-pro-002閉源Google	1,266	±4	23,856	—
4	—	gpt-4o-2024-05-13閉源OpenAI	1,262	±3	108,575	—
5	—	o1-mini閉源OpenAI	1,257	±5	28,953	—
6	—	gemini-advanced-0514閉源Google	1,253	±4	52,219	—
7	—	llama-3.1-405b-instruct-fp8開源Meta	1,252	±4	48,804	—
8	—	llama-3.1-405b-instruct-bf16開源Meta	1,251	±5	14,611	—
9	—	grok-2-2024-08-13閉源xAI	1,249	±3	47,908	—
10	—	gpt-4o-2024-08-06閉源OpenAI	1,248	±4	40,499	—
11	—	yi-lightning閉源01 AI	1,244	±4	27,114	—
12	—	gpt-4-turbo-2024-04-09閉源OpenAI	1,241	±3	101,865	—
13	—	gemini-1.5-pro-001閉源Google	1,239	±3	82,621	—
14	—	claude-3-opus-20240229閉源Anthropic	1,238	±2	183,545	—
15	—	gpt-4-1106-preview閉源OpenAI	1,234	±3	103,870	—
16	—	glm-4-plus閉源Zhipu AI	1,233	±5	25,601	—
17	—	gpt-4o-mini-2024-07-18閉源OpenAI	1,231	±4	48,407	—
18	▲ 1	mistral-large-2407開源Mistral · 另有 1 個檔位	1,230	±3	45,046	—
19	▼ 1	gpt-4-0125-preview閉源OpenAI	1,230	±3	97,202	—
20	—	qwen-max-0919開源Alibaba	1,229	±4	17,642	—
21	—	gemini-1.5-flash-002閉源Google	1,228	±5	18,112	—
22	—	deepseek-v2.5開源DeepSeek	1,223	±4	22,918	—
23	—	grok-2-mini-2024-08-13閉源xAI	1,223	±4	39,214	—
24	—	qwen2.5-72b-instruct開源Alibaba	1,222	±4	23,112	—
25	—	athene-70b-0725開源NexusFlow	1,221	±5	20,697	—
26	—	llama-3.1-nemotron-70b-instruct開源NVIDIA	1,214	±7	7,263	—
27	—	llama-3.1-70b-instruct開源Meta	1,212	±4	44,687	—
28	▲ 1	gemma-2-27b-it開源Google	1,205	±4	54,961	—
29	▲ 3	llama-3.1-nemotron-51b-instruct開源NVIDIA	1,204	±7	3,943	—
30	—	jamba-1.5-large開源AI21 Labs	1,203	±6	9,186	—
31	—	gemini-1.5-flash-001閉源Google	1,203	±3	65,794	—
32	▲ 1	gpt-4-0314閉源OpenAI	1,200	±3	55,966	—
33	▲ 1	gemma-2-9b-it-simpo開源Princeton	1,198	±6	10,613	—
34	▲ 1	claude-3-sonnet-20240229閉源Anthropic	1,197	±2	113,045	—
35	▲ 1	command-r-plus-08-2024開源Cohere	1,197	±5	10,608	—
36	▲ 1	nemotron-4-340b-instruct開源NVIDIA	1,196	±4	20,641	—
37	▲ 1	llama-3-70b-instruct開源Meta	1,194	±2	163,858	—
38	▲ 2	glm-4-0520閉源Zhipu AI	1,187	±6	10,228	—
39	▲ 2	gpt-4-0613閉源OpenAI	1,186	±3	91,646	—
40	▲ 2	command-r-plus開源Cohere	1,183	±3	80,879	—
41	▲ 2	deepseek-coder-v2開源DeepSeek	1,181	±5	15,796	—
42	▲ 2	qwen2-72b-instruct開源Alibaba	1,180	±4	38,980	—
43	▲ 4	gemma-2-9b-it開源Google	1,177	±3	39,510	—
44	▲ 1	claude-3-haiku-20240307閉源Anthropic	1,177	±3	122,422	—
45	▲ 1	gemini-1.5-flash-8b-001閉源Google	1,177	±5	19,267	—
46	▲ 2	command-r-08-2024開源Cohere	1,176	±5	10,924	—
47	▲ 2	jamba-1.5-mini開源AI21 Labs	1,159	±7	9,296	—
48	新	ministral-8b-2410開源Mistral	1,156	±12	2,333	—
49	▲ 1	qwen1.5-110b-chat開源Alibaba	1,155	±4	27,472	—
50	▲ 1	reka-flash-21b-20240226-online閉源Reka AI	1,151	±5	16,023	—
51	▲ 1	mixtral-8x22b-instruct-v0.1開源Mistral	1,149	±3	53,814	—
52	新	hunyuan-standard-256k閉源Tencent	1,149	±9	2,860	—
53	—	command-r開源Cohere	1,148	±3	56,400	—
54	—	qwen1.5-72b-chat開源Alibaba	1,148	±3	40,636	—
55	—	gemini-pro-dev-api閉源Google	1,147	±4	18,787	—
56	—	llama-3-8b-instruct開源Meta	1,142	±3	109,312	—
57	—	reka-flash-21b-20240226閉源Reka AI	1,141	±4	25,802	—
58	—	mistral-medium閉源Mistral	1,141	±3	35,537	—
59	—	gpt-3.5-turbo-0125閉源OpenAI · 另有 1 個檔位	1,137	±3	68,876	—
60	—	yi-1.5-34b-chat開源01 AI	1,132	±3	25,150	—
61	—	gemini-pro閉源Google	1,131	±7	6,561	—
62	—	zephyr-orpo-141b-A35b-v0.1開源HuggingFace	1,130	±7	4,849	—
63	—	llama-3.1-8b-instruct開源Meta	1,130	±3	42,983	—
64	—	qwen1.5-32b-chat開源Alibaba	1,127	±4	22,753	—
65	—	phi-3-medium-4k-instruct開源Microsoft	1,118	±4	26,143	—
66	—	internlm2_5-20b-chat開源InternLM	1,115	±6	10,709	—
67	—	mixtral-8x7b-instruct-v0.1開源Mistral	1,114	±0	76,142	—
68	—	dbrx-instruct-preview開源Databricks	1,113	±3	33,718	—
69	—	gemma-2-2b-it開源Google	1,112	±4	31,041	—
70	—	qwen1.5-14b-chat開源Alibaba	1,107	±4	18,666	—
71	—	wizardlm-70b開源Microsoft	1,105	±7	8,382	—
72	—	snowflake-arctic-instruct開源Snowflake	1,103	±4	34,163	—
73	—	yi-34b-chat開源01 AI	1,101	±4	15,919	—
74	—	openchat-3.5-0106開源OpenChat · 另有 1 個檔位	1,100	±5	12,971	—
75	—	phi-3-small-8k-instruct開源Microsoft	1,098	±5	18,499	—
76	—	gemma-1.1-7b-it開源Google	1,098	±4	25,074	—
77	—	llama-3.2-3b-instruct開源Meta	1,093	±6	8,454	—
78	—	tulu-2-dpo-70b開源AllenAI/UW	1,092	±7	6,662	—
79	—	vicuna-33b開源LMSYS	1,091	±4	22,941	—
80	—	starling-lm-7b-beta開源Nexusflow	1,091	±4	16,665	—
81	—	deepseek-llm-67b-chat開源DeepSeek	1,091	±7	4,984	—
82	—	openhermes-2.5-mistral-7b開源NousResearch	1,090	±8	5,090	—
83	—	starling-lm-7b-alpha開源UC Berkeley	1,083	±6	10,414	—
84	—	llama-2-70b-chat開源Meta	1,081	±3	39,617	—
85	—	llama2-70b-steerlm-chat開源NVIDIA	1,074	±10	3,638	—
86	—	nous-hermes-2-mixtral-8x7b-dpo開源NousResearch	1,071	±9	3,837	—
87	—	dolphin-2.2.1-mistral-7b開源Cognitive Computations	1,068	±14	1,713	—
88	—	phi-3-mini-4k-instruct-june-2024開源Microsoft	1,068	±6	12,874	—
89	—	mistral-7b-instruct-v0.2開源Mistral	1,067	±4	20,060	—
90	—	solar-10.7b-instruct-v1.0開源Upstage AI	1,066	±8	4,288	—
91	—	qwen1.5-7b-chat開源Alibaba	1,065	±7	4,862	—
92	—	wizardlm-13b開源Microsoft	1,063	±8	7,183	—
93	—	mpt-30b-chat開源MosaicML	1,060	±12	2,649	—
94	—	falcon-180b-chat開源TII	1,059	±16	1,326	—
95	—	vicuna-13b開源LMSYS	1,059	±5	19,782	—
96	—	qwen-14b-chat開源Alibaba	1,057	±10	5,069	—
97	—	phi-3-mini-4k-instruct開源Microsoft	1,055	±5	21,118	—
98	—	llama-2-13b-chat開源Meta	1,054	±5	19,730	—
99	—	zephyr-7b-beta開源HuggingFace	1,049	±7	11,325	—
100	—	phi-3-mini-128k-instruct開源Microsoft	1,049	±4	21,616	—
101	▲ 1	gemma-7b-it開源Google	1,047	±6	9,177	—
102	▼ 1	codellama-34b-instruct開源Meta	1,047	±7	7,509	—
103	—	zephyr-7b-alpha開源HuggingFace	1,046	±16	1,813	—
104	—	llama-3.2-1b-instruct開源Meta	1,044	±7	8,568	—
105	—	palm-2閉源Google	1,043	±5	8,743	—
106	—	guanaco-33b開源UW	1,040	±12	2,997	—
107	—	gemma-1.1-2b-it開源Google	1,039	±6	11,349	—
108	—	codellama-70b-instruct開源Meta	1,036	±15	1,190	—
109	—	stripedhyena-nous-7b開源Together AI	1,034	±9	5,272	—
110	—	mistral-7b-instruct開源Mistral	1,027	±7	9,139	—
111	—	llama-2-7b-chat開源Meta	1,024	±6	14,551	—
112	—	vicuna-7b開源LMSYS	1,022	±7	7,017	—
113	—	qwen1.5-4b-chat開源Alibaba	1,015	±7	7,812	—
114	—	gemma-2b-it開源Google	1,007	±7	4,913	—
115	—	olmo-7b-instruct開源Ai2	994	±7	6,495	—
116	—	koala-13b開源UC Berkeley	980	±7	7,033	—
117	—	chatglm3-6b開源Tsinghua	975	±9	4,765	—
118	—	gpt4all-13b-snoozy開源Nomic AI	972	±14	1,786	—
119	—	alpaca-13b開源Stanford	968	±8	5,872	—
120	—	mpt-7b-chat開源MosaicML	967	±9	4,012	—
121	—	RWKV-4-Raven-14B開源RWKV	949	±10	4,935	—
122	—	chatglm2-6b開源Tsinghua	943	±10	2,708	—
123	—	oasst-pythia-12b開源OpenAssistant	932	±9	6,382	—
124	—	fastchat-t5-3b開源LMSYS	896	±9	4,300	—
125	—	chatglm-6b開源Tsinghua	889	±9	4,993	—
126	—	dolly-v2-12b開源Databricks	882	±11	3,485	—
127	—	llama-13b開源Meta	868	±13	2,445	—
128	—	stablelm-tuned-alpha-7b開源Stability AI	859	±11	3,338	—

其餘 78 個模型

智力進步

每家廠商當時最強的模型在 Artificial Analysis 智力指數上的分數，按模型釋出的月份畫。

漲得最多 · Anthropic +57.2目前最高 · Anthropic 64.6最高與最低相差 18.1 分時間範圍 2025年11月 – 2026年11月

64.6Anthropic60.0Google56.9OpenAI50.1Meta46.5Kimi