模型榜单

排名更新时间: 2025-05-18

排名(UB)	排名(StyleCtrl)	模型名	分数	置信区间	票数	服务商	协议
1	1	gemini-2.5-pro-preview-05-06	1446	+8/-9	4500	Google	Proprietary
2	1	o3-2025-04-16	1413	+8/-7	6689	OpenAI	Proprietary
2	2	chatgpt-4o-latest-20250326	1408	+6/-6	10290	OpenAI	Proprietary
2	5	grok-3-preview-02-24	1403	+4/-4	14843	xAI	Proprietary
3	3	gpt-4.5-preview-2025-02-27	1398	+4/-5	15275	OpenAI	Proprietary
4	5	gemini-2.5-flash-preview-04-17	1394	+7/-7	5959	Google	Proprietary
7	5	deepseek-v3-0324	1373	+7/-5	8753	DeepSeek	MIT
7	5	gpt-4.1-2025-04-14	1366	+7/-8	5102	OpenAI	Proprietary
8	8	deepseek-r1	1358	+4/-4	18493	DeepSeek	MIT
8	15	gemini-2.0-flash-001	1355	+4/-3	24913	Google	Proprietary
8	13	hunyuan-turbos-20250416	1355	+9/-10	3699	Tencent	Proprietary
8	5	o4-mini-2025-04-16	1351	+10/-7	5083	OpenAI	Proprietary
9	7	o1-2024-12-17	1350	+4/-4	29036	OpenAI	Proprietary
9	14	qwen3-235b-a22b	1343	+11/-9	3611	Alibaba	Apache 2.0
11	15	gemma-3-27b-it	1341	+5/-4	12343	Google	Gemma
12	15	qwen2.5-max	1341	+4/-3	23180	Alibaba	Proprietary
14	12	o1-preview	1335	+3/-3	33171	OpenAI	Proprietary
17	24	gemma-3-12b-it	1321	+11/-11	3016	Google	Gemma
18	15	o3-mini-high	1325	+4/-4	19410	OpenAI	Proprietary
18	13	gpt-4.1-mini-2025-04-14	1322	+6/-7	4950	OpenAI	Proprietary
18	21	deepseek-v3	1318	+4/-3	22838	DeepSeek	DeepSeek
19	27	qwq-32b	1313	+6/-5	9946	Alibaba	Apache 2.0
19	23	gemini-2.0-flash-lite-preview-02-05	1312	+3/-3	25020	Google	Proprietary
19	27	glm-4-plus-0111	1311	+6/-7	6024	Zhipu	Proprietary
19	24	qwen-plus-0125	1310	+5/-6	6058	Alibaba	Proprietary
21	24	command-a-03-2025	1306	+4/-5	10513	Cohere	CC-BY-NC-4.0
21	29	step-2-16k-exp-202412	1305	+8/-7	5125	StepFun	Proprietary
21	23	hunyuan-turbos-20250226	1303	+11/-9	2449	Tencent	Proprietary
22	23	o3-mini	1305	+4/-3	24918	OpenAI	Proprietary
24	29	o1-mini	1304	+2/-3	54953	OpenAI	Proprietary
24	24	gemini-1.5-pro-002	1302	+2/-2	58645	Google	Proprietary
24	10	claude-3-7-sonnet-20250219-thinking	1300	+5/-5	12038	Anthropic	Proprietary
24	30	llama-3.3-nemotron-49b-super-v1	1297	+9/-10	2366	Nvidia	Nvidia
24	24	hunyuan-turbo-0110	1296	+11/-10	2512	Tencent	Proprietary
31	13	claude-3-7-sonnet-20250219	1290	+6/-5	17387	Anthropic	Proprietary
33	35	grok-2-2024-08-13	1288	+2/-2	67079	xAI	Proprietary
33	38	yi-lightning	1287	+3/-4	28968	01 AI	Proprietary
33	26	gpt-4o-2024-05-13	1285	+2/-1	117741	OpenAI	Proprietary
33	47	qwen2.5-plus-1127	1282	+5/-4	10711	Alibaba	Proprietary
33	54	gemma-3-4b-it	1275	+12/-10	3326	Google	Gemma
34	42	deepseek-v2.5-1210	1279	+7/-6	7242	DeepSeek	DeepSeek
36	15	claude-3-5-sonnet-20241022	1283	+2/-3	65435	Anthropic	Proprietary
38	38	hunyuan-large-2025-02-10	1272	+11/-6	3858	Tencent	Proprietary
40	50	athene-v2-chat	1275	+3/-3	26077	NexusFlow	NexusFlow
40	40	gpt-4.1-nano-2025-04-14	1271	+7/-6	5121	OpenAI	Proprietary
41	45	glm-4-plus	1274	+3/-3	27787	Zhipu AI	Proprietary
41	47	gpt-4o-mini-2024-07-18	1272	+2/-2	71363	OpenAI	Proprietary
41	50	gemini-1.5-flash-002	1271	+2/-3	37025	Google	Proprietary
41	31	llama-4-maverick-17b-128e-instruct	1270	+6/-6	7798	Meta	Llama 4
41	63	llama-3.1-nemotron-70b-instruct	1269	+7/-6	7579	Nvidia	Llama 3.1
42	30	llama-3.1-405b-instruct-bf16	1269	+3/-2	43799	Meta	Llama 3.1 Community
44	28	claude-3-5-sonnet-20240620	1268	+2/-2	86162	Anthropic	Proprietary
44	30	gemini-advanced-0514	1267	+3/-2	52136	Google	Proprietary
44	47	hunyuan-standard-2025-02-10	1261	+9/-8	4015	Tencent	Proprietary
45	31	llama-3.1-405b-instruct-fp8	1267	+2/-2	63030	Meta	Llama 3.1 Community
45	62	grok-2-mini-2024-08-13	1266	+2/-2	55434	xAI	Proprietary
46	34	gpt-4o-2024-08-06	1265	+2/-3	47970	OpenAI	Proprietary
46	49	qwen-max-0919	1263	+4/-5	17432	Alibaba	Qwen
56	44	gemini-1.5-pro-001	1260	+2/-2	82431	Google	Proprietary
56	58	deepseek-v2.5	1258	+4/-4	26339	DeepSeek	DeepSeek
57	63	qwen2.5-72b-instruct	1257	+3/-2	41517	Alibaba	Qwen
57	47	llama-3.3-70b-instruct	1257	+3/-3	38088	Meta	Llama-3.3
57	42	gpt-4-turbo-2024-04-09	1256	+2/-2	102145	OpenAI	Proprietary
59	50	mistral-large-2407	1252	+3/-2	48217	Mistral	Mistral Research
63	63	athene-70b-0725	1250	+3/-4	20578	NexusFlow	CC-BY-NC-4.0
63	63	mistral-large-2411	1249	+4/-3	29643	Mistral	MRL
64	47	gpt-4-1106-preview	1250	+2/-2	103749	OpenAI	Proprietary
64	69	llama-3.1-70b-instruct	1248	+2/-2	58637	Meta	Llama 3.1 Community
64	69	llama-3.1-tulu-3-70b	1244	+8/-9	3009	Ai2	Llama 3.1
65	45	claude-3-opus-20240229	1247	+1/-2	202651	Anthropic	Proprietary
65	69	amazon-nova-pro-v1.0	1245	+3/-3	25794	Amazon	Proprietary
66	50	gpt-4-0125-preview	1245	+2/-2	97080	OpenAI	Proprietary
72	47	claude-3-5-haiku-20241022	1237	+3/-3	36354	Anthropic	Propretary
72	69	reka-core-20240904	1235	+5/-5	7946	Reka AI	Proprietary
75	72	gemini-1.5-flash-001	1227	+2/-3	65662	Google	Proprietary
75	71	jamba-1.5-large	1222	+6/-5	9125	AI21 Labs	Jamba Open
75	79	qwen2.5-coder-32b-instruct	1217	+7/-8	5733	Alibaba	Apache 2.0
76	72	gemma-2-27b-it	1220	+2/-2	79527	Google	Gemma license
76	79	mistral-small-24b-instruct-2501	1218	+4/-4	15323	Mistral	Apache 2.0
76	87	amazon-nova-lite-v1.0	1217	+3/-3	20652	Amazon	Proprietary
76	74	gemma-2-9b-it-simpo	1216	+5/-4	10549	Princeton	MIT
76	75	command-r-plus-08-2024	1215	+5/-5	10537	Cohere	CC-BY-NC-4.0
76	70	llama-3.1-nemotron-51b-instruct	1212	+8/-10	3890	Nvidia	Llama 3.1
78	88	gemini-1.5-flash-8b-001	1213	+3/-3	37695	Google	Proprietary
78	87	olmo-2-0325-32b-instruct	1206	+9/-9	3183	Allen AI	Apache-2.0
80	86	c4ai-aya-expanse-32b	1209	+3/-3	28757	Cohere	CC-BY-NC-4.0
80	77	nemotron-4-340b-instruct	1209	+3/-4	20610	Nvidia	NVIDIA Open Model
80	82	glm-4-0520	1207	+5/-5	10224	Zhipu AI	Proprietary
81	77	reka-flash-20240904	1206	+5/-6	8134	Reka AI	Proprietary
83	79	llama-3-70b-instruct	1207	+2/-2	163642	Meta	Llama 3 Community
83	91	phi-4	1205	+4/-3	25221	Microsoft	MIT
87	77	claude-3-sonnet-20240229	1201	+2/-2	113071	Anthropic	Proprietary
87	100	amazon-nova-micro-v1.0	1198	+5/-3	20660	Amazon	Proprietary
90	101	hunyuan-standard-256k	1189	+11/-9	2900	Tencent	Proprietary
94	88	gemma-2-9b-it	1192	+2/-2	57201	Google	Gemma license
94	87	command-r-plus	1190	+2/-2	80856	Cohere	CC-BY-NC-4.0
94	102	llama-3.1-tulu-3-8b	1185	+8/-10	3075	Ai2	Llama 3.1
94	100	ministral-8b-2410	1182	+8/-7	5109	Mistral	MRL
95	87	qwen2-72b-instruct	1187	+2/-2	38870	Alibaba	Qianwen LICENSE
95	75	gpt-4-0314	1186	+3/-3	55967	OpenAI	Proprietary
96	102	c4ai-aya-expanse-8b	1180	+6/-5	10395	Cohere	CC-BY-NC-4.0
97	91	command-r-08-2024	1180	+4/-5	10849	Cohere	CC-BY-NC-4.0
97	87	deepseek-coder-v2	1178	+5/-5	15756	DeepSeek AI	DeepSeek License
98	92	claude-3-haiku-20240307	1179	+2/-2	122315	Anthropic	Proprietary
98	101	jamba-1.5-mini	1176	+5/-5	9273	AI21 Labs	Jamba Open
99	117	llama-3.1-8b-instruct	1176	+2/-3	52576	Meta	Llama 3.1 Community
107	86	gpt-4-0613	1163	+2/-2	91628	OpenAI	Proprietary
107	102	qwen1.5-110b-chat	1161	+3/-4	27438	Alibaba	Qianwen LICENSE
107	117	yi-1.5-34b-chat	1157	+4/-3	25136	01 AI	Apache-2.0
107	103	reka-flash-21b-20240226-online	1156	+6/-4	16027	Reka AI	Proprietary
107	132	qwq-32b-preview	1153	+11/-8	3412	Alibaba	Apache 2.0
108	101	mistral-large-2402	1157	+2/-3	64924	Mistral	Proprietary
109	109	llama-3-8b-instruct	1152	+2/-3	109093	Meta	Llama 3 Community
111	124	internlm2_5-20b-chat	1149	+3/-5	10596	InternLM	Other
111	111	granite-3.1-8b-instruct	1143	+10/-8	3290	IBM	Apache 2.0
112	105	command-r	1149	+2/-3	56392	Cohere	CC-BY-NC-4.0
112	105	mixtral-8x22b-instruct-v0.1	1148	+2/-3	53761	Mistral	Apache 2.0
112	110	mistral-medium	1148	+3/-3	35566	Mistral	Proprietary
112	106	qwen1.5-72b-chat	1147	+3/-3	40665	Alibaba	Qianwen LICENSE
112	108	reka-flash-21b-20240226	1147	+3/-4	25807	Reka AI	Proprietary
113	124	gemma-2-2b-it	1144	+3/-2	48901	Google	Gemma license
121	105	gemini-pro-dev-api	1131	+4/-4	18803	Google	Proprietary
122	114	zephyr-orpo-141b-A35b-v0.1	1127	+7/-10	4853	HuggingFace	Apache 2.0
122	119	qwen1.5-32b-chat	1125	+3/-4	22762	Alibaba	Qianwen LICENSE
122	124	granite-3.1-2b-instruct	1119	+8/-11	3380	IBM	Apache 2.0
123	123	phi-3-medium-4k-instruct	1123	+3/-3	26112	Microsoft	MIT
123	134	starling-lm-7b-beta	1119	+4/-4	16674	Nexusflow	Apache-2.0
125	114	gemini-pro	1111	+7/-6	6557	Google	Proprietary
126	124	mixtral-8x7b-instruct-v0.1	1114	+2/-2	76134	Mistral	Apache 2.0
126	129	yi-34b-chat	1111	+5/-5	15918	01 AI	Yi License
127	126	qwen1.5-14b-chat	1109	+4/-3	18685	Alibaba	Qianwen LICENSE
127	127	wizardlm-70b	1106	+7/-6	8383	Microsoft	Llama 2 Community
128	114	gpt-3.5-turbo-0125	1106	+2/-2	68871	OpenAI	Proprietary
128	132	llama-3.2-3b-instruct	1103	+6/-6	8394	Meta	Llama 3.2
129	124	dbrx-instruct-preview	1103	+3/-3	33743	Databricks	DBRX LICENSE
129	131	phi-3-small-8k-instruct	1102	+4/-4	18477	Microsoft	MIT
129	132	tulu-2-dpo-70b	1099	+7/-7	6658	AllenAI/UW	AI2 ImpACT Low-risk
135	124	granite-3.0-8b-instruct	1093	+6/-7	7002	IBM	Apache 2.0
136	129	openchat-3.5-0106	1092	+5/-5	12986	OpenChat	Apache-2.0
137	143	llama-2-70b-chat	1093	+3/-4	39602	Meta	Llama 2 Community
137	135	vicuna-33b	1091	+3/-4	22937	LMSYS	Non-commercial
137	129	snowflake-arctic-instruct	1090	+2/-3	34181	Snowflake	Apache 2.0
137	139	starling-lm-7b-alpha	1088	+5/-6	10414	UC Berkeley	CC-BY-NC-4.0
137	146	nous-hermes-2-mixtral-8x7b-dpo	1084	+8/-8	3837	NousResearch	Apache-2.0
138	143	llama2-70b-steerlm-chat	1080	+11/-10	3636	Nvidia	Llama 2 Community
139	131	gemma-1.1-7b-it	1084	+3/-3	25063	Google	Gemma license
143	132	deepseek-llm-67b-chat	1077	+7/-9	4988	DeepSeek AI	DeepSeek License
143	132	openchat-3.5	1076	+6/-7	8107	OpenChat	Apache-2.0
143	139	granite-3.0-2b-instruct	1074	+8/-8	7187	IBM	Apache 2.0
144	134	openhermes-2.5-mistral-7b	1074	+7/-7	5086	NousResearch	Apache-2.0
145	150	phi-3-mini-4k-instruct-june-2024	1071	+7/-5	12800	Microsoft	MIT
145	150	qwen1.5-7b-chat	1070	+7/-7	4873	Alibaba	Qianwen LICENSE
145	146	dolphin-2.2.1-mistral-7b	1062	+14/-11	1713	Cognitive Computations	Apache-2.0
146	150	mistral-7b-instruct-v0.2	1072	+3/-4	20063	Mistral	Apache-2.0
146	126	gpt-3.5-turbo-1106	1068	+4/-5	17035	OpenAI	Proprietary
146	154	phi-3-mini-4k-instruct	1066	+4/-3	21092	Microsoft	MIT
146	149	solar-10.7b-instruct-v1.0	1062	+8/-8	4286	Upstage AI	CC-BY-NC-4.0
150	154	llama-2-13b-chat	1063	+4/-3	19713	Meta	Llama 2 Community
151	150	wizardlm-13b	1059	+7/-7	7175	Microsoft	Llama 2 Community
156	160	llama-3.2-1b-instruct	1054	+7/-6	8518	Meta	Llama 3.2
157	158	zephyr-7b-beta	1053	+4/-6	11321	HuggingFace	MIT
157	154	smollm2-1.7b-instruct	1046	+10/-12	2376	HuggingFace	Apache 2.0
157	150	mpt-30b-chat	1045	+11/-10	2644	MosaicML	CC-BY-NC-SA-4.0
157	159	codellama-70b-instruct	1041	+16/-12	1192	Meta	Llama 2 Community
159	157	zephyr-7b-alpha	1040	+11/-13	1812	HuggingFace	MIT
160	158	codellama-34b-instruct	1043	+6/-7	7508	Meta	Llama 2 Community
160	150	falcon-180b-chat	1034	+15/-16	1327	TII	Falcon-180B TII License
161	152	vicuna-13b	1042	+5/-4	19773	LMSYS	Llama 2 Community
162	158	gemma-7b-it	1037	+5/-7	9176	Google	Gemma license
162	173	llama-2-7b-chat	1037	+5/-4	14532	Meta	Llama 2 Community
162	158	phi-3-mini-128k-instruct	1037	+4/-4	21625	Microsoft	MIT
162	150	qwen-14b-chat	1035	+7/-9	5067	Alibaba	Qianwen LICENSE
162	162	guanaco-33b	1033	+9/-9	2995	UW	Non-commercial
170	163	gemma-1.1-2b-it	1021	+6/-5	11349	Google	Gemma license
172	167	stripedhyena-nous-7b	1017	+8/-8	5277	Together AI	Apache 2.0
173	180	olmo-7b-instruct	1015	+6/-7	6501	Allen AI	Apache-2.0
175	172	mistral-7b-instruct	1008	+7/-8	9144	Mistral	Apache 2.0
175	173	vicuna-7b	1005	+8/-8	7015	LMSYS	Llama 2 Community
175	162	palm-2	1003	+7/-5	8711	Google	Proprietary
178	178	gemma-2b-it	990	+9/-6	4918	Google	Gemma license
180	175	qwen1.5-4b-chat	989	+5/-7	7819	Alibaba	Qianwen LICENSE
182	181	koala-13b	965	+6/-8	7020	UC Berkeley	Non-commercial
182	182	chatglm3-6b	955	+8/-10	4762	Tsinghua	Apache-2.0
183	182	gpt4all-13b-snoozy	932	+13/-14	1788	Nomic AI	Non-commercial
184	182	mpt-7b-chat	928	+9/-10	3997	MosaicML	CC-BY-NC-SA-4.0
184	187	chatglm2-6b	924	+12/-13	2712	Tsinghua	Apache-2.0
184	185	RWKV-4-Raven-14B	922	+10/-10	4920	RWKV	Apache 2.0
188	182	alpaca-13b	901	+8/-7	5864	Stanford	Non-commercial
188	188	oasst-pythia-12b	893	+7/-7	6368	OpenAssistant	Apache 2.0
189	190	chatglm-6b	879	+9/-9	4983	Tsinghua	Non-commercial
190	190	fastchat-t5-3b	868	+9/-9	4288	LMSYS	Apache 2.0
192	193	stablelm-tuned-alpha-7b	840	+7/-10	3336	Stability AI	CC-BY-NC-SA-4.0
192	190	dolly-v2-12b	822	+10/-13	3480	Databricks	MIT
193	191	llama-13b	800	+11/-13	2446	Meta	Non-commercial

说明

• 排名(UB)：基于 Bradley-Terry 模型的上界排名

• 排名(StyleCtrl)：考虑对话风格的样式控制排名

• 置信区间：模型表现的置信区间

• 分数：基于模型性能的竞技场得分

• 票数：模型获得的投票数量

• 服务商：提供模型的机构或公司

• 协议：模型使用的许可协议

数据来源

• 数据来自 lmarena.ai