モデルランキング

このドキュメントはAIによって中国語から翻訳されており、まだレビューされていません。

これはChatbot Arena (lmarena.ai)のデータに基づくランキングで、自動化プロセスにより生成されています。

データ更新時刻: 2025-08-11 11:44:05 UTC / 2025-08-11 19:44:05 CST (北京時間)

ランキング内のモデル名をクリックすると、詳細情報や試用ページにジャンプできます。

ランキング

順位(UB)	順位(StyleCtrl)	モデル名	スコア	信頼区間	投票数	プロバイダ	ライセンス	ナレッジカットオフ
1	1	Gemini-2.5-Pro	1470	+5/-5	26,019	Google	Proprietary	nan
2	2	Gemini-2.5-Pro-Preview-05-06	1446	+6/-6	13,715	Google	Proprietary	nan
3	2	GLM-4.5	1434	+9/-9	4,112	Z.ai	MIT	nan
4	2	Grok-4-0709	1434	+6/-6	13,058	xAI	Proprietary	nan
5	3	ChatGPT-4o-latest (2025-03-26)	1429	+4/-4	30,777	OpenAI	Proprietary	nan
6	3	o3-2025-04-16	1428	+4/-4	32,033	OpenAI	Proprietary	nan
7	3	Qwen3-235B-A22B-Instruct-2507	1427	+9/-9	4,154	Alibaba	Apache 2.0	nan
8	3	DeepSeek-R1-0528	1427	+5/-5	18,284	DeepSeek	MIT	nan
9	4	Grok-3-Preview-02-24	1423	+4/-4	31,757	xAI	Proprietary	nan
10	8	Llama-4-Maverick-03-26-Experimental	1416	+4/-4	26,604	Meta	nan	nan
11	8	GPT-4.5-Preview	1415	+5/-5	15,271	OpenAI	Proprietary	nan
12	7	Qwen3-235B-A22B-Thinking-2507	1413	+9/-9	3,715	Alibaba	Apache 2.0	nan
13	8	chocolate (Early Grok-3)	1412	+6/-6	13,837	xAI	Proprietary	nan
14	10	Gemini-2.5-Flash	1411	+4/-4	31,359	Google	Proprietary	nan
15	15	Gemini-2.0-Flash-Thinking-Exp-01-21	1397	+4/-4	27,552	Google	Proprietary	nan
16	15	Gemini-2.0-Pro-Exp-02-05	1397	+5/-5	20,120	Google	Proprietary	nan
17	15	Gemini-2.5-Flash-Preview-04-17	1396	+5/-5	18,655	Google	Proprietary	nan
18	15	GLM-4.5-Air	1393	+9/-9	4,306	Z.ai	MIT	nan
19	15	Qwen3-235B-A22B-no-thinking	1391	+5/-5	24,372	Alibaba	Apache 2.0	nan
20	15	Gemini-Exp-1206	1389	+4/-4	23,657	Google	Proprietary	nan
21	15	ChatGPT-4o-latest (2025-01-29)	1389	+4/-4	23,858	OpenAI	Proprietary	nan
22	19	ChatGPT-4o-latest (2024-11-20)	1381	+3/-3	40,509	OpenAI	Proprietary	nan
23	18	kimi-k2-0711-preview	1380	+6/-6	11,676	Moonshot	Modified MIT	nan
24	19	GPT-4.1-2025-04-14	1380	+5/-5	24,834	OpenAI	Proprietary	nan
25	16	Qwen3-30B-A3B-Instruct-2507	1380	+12/-12	2,258	Alibaba	Apache 极2.0	nan
26	22	Gemini-Exp-1121	1379	+5/-5	17,328	Google	Proprietary	nan
27	22	Gemini-2.0-Flash-Thinking-Exp-1219	1378	+5/-5	16,963	Google	Proprietary	nan
28	22	Hunyuan-Turbos-20250416	1376	+6/-6	11,657	Tencent	Proprietary	nan
29	22	DeepSeek-V3-0324	1376	+4/-4	27,391	DeepSeek	MIT	nan
30	22	Claude Opus 4 (thinking-16k)	1373	+5/-5	17,970	Anthropic	Proprietary	nan
31	23	DeepSeek-R1	1372	+4/-4	19,430	DeepSeek	MIT	nan
32	24	Gemini-2.0-Flash-Exp	1370	+4/-4	22,500	Google	Proprietary	nan
33	24	Mistral Medium 3	1370	+5/-5	28,010	Mistral	Proprietary	nan
34	24	Gemini-Exp-1114	1368	+5/-5	17,088	Google	Proprietary	nan
35	28	Qwen2.5-Max	1366	+3/-3	35,457	Alibaba	Proprietary	nan
36	27	Claude Opus 4 (20250514)	1366	+5/-5	26,141	Anthropic	Proprietary	nan
37	28	Qwen3-235B-A22B	1365	+5/-5	20,489	Alibaba	Apache 2.0	nan
38	29	o1-2024-12-17	1365	+4/-4	29,038	OpenAI	Proprietary	nan
39	32	Gemini-2.0-Flash-001	1362	+3/-3	41,036	Google	Proprietary	nan
40	32	o4-mini-2025-04-16	1362	+5/-5	24,472	OpenAI	Proprietary	nan
41	32	Grok-3-mini-high	1361	+6/-6	10,194	xAI	Proprietary	nan
42	32	Grok-3-Mini-beta	1360	+5/-5	16,845	xAI	Proprietary	nan
43	33	Qwen3-Coder-480B-A35B-Instruct	1357	+7/-7	6,012	Alibaba	Apache 2.0	nan
44	36	Gemma-3-27B-it	1357	+4/-4	32,176	Google	Gemma	nan
45	39	ChatGPT-4o-latest (2024-09-03)	1354	+4/-4	48,669	OpenAI	Proprietary	2023/10
46	39	Minimax-M1	1354	+5/-5	17,524	MiniMax	Apache 2.0	nan
47	40	o1-preview 极	1353	+4/-4	33,177	OpenAI	Proprietary	2023/10
48	41	Claude Sonnet 4 (thinking-32k)	1350	+5/-5	17,124	Anthropic	Proprietary	nan
49	44	Qwen3-32B	1341	+9/-9	4,074	Alibaba	Apache 2.0	nan
50	46	Nvidia-Llama-3.3-Nemotron-Super-49B-v1.5	1339	+10/-10	3,106	Nvidia	Nvidia Open	nan
51	49	o3-mini-high	1338	+5/-5	19,404	OpenAI	Proprietary	nan
52	49	Claude Sonnet 4 (20250514)	1338	+5/-5	22,629	Anthropic	Proprietary	nan
53	49	GPT-4.1-mini-2025-04-14	1336	+5/-5	23,892	OpenAI	Proprietary	nan
54	49	Mistral-Small-2506	1335	+6/-6	9,147	Mistral	Apache 2.0	nan
55	49	Gemma-3-12B-it	1335	+9/-9	3,976	Google	Gemma	nan
56	49	ChatGPT-4o-latest (2024-08-08)	1334	+3/-3	35,503	OpenAI	Proprietary	2023/10
57	49	DeepSeek-V3	1333	+4/-4	22,841	DeepSeek	DeepSeek	nan
58	49	GLM-4-Plus-0111	1332	+8/-8	6,028	Zhipu	Proprietary	nan
59	49	QwQ-32B	1331	+4/-4	22,178	Alibaba	Apache 2.0	nan
60	49	Gemini-2.0-Flash-Lite	1330	+4/-4	26,104	Google	Proprietary	nan
61	49	Qwen-Plus-0125	1327	+8/-8	6,055	Alibaba	Proprietary	nan
62	54	Command A (03-2025)	1326	+4/-4	30,944	Cohere	CC-BY-NC-4.0	nan
63	57	Amazon-Nova-Experimental-Chat-05-14	1322	+5/-5	13,900	Amazon	Proprietary	nan
64	55	Step-2-16K-Exp	1320	+8/-8	5,126	StepFun	Proprietary	nan
65	52	Llama-3.1-Nemotron-Ultra-253B-v1	1320	+11/-11	2,656	Nvidia	Nvidia Open Model	nan
66	60	Qwen3-30B-A3B	1320	+5/-5	20,360	Alibaba	Apache 2.0	nan
67	61	Gemini-1.5-Pro-002	1320	+3/-3	58,645	Google	Proprietary	nan
68	53	Hunyuan-TurboS-20250226	1320	+10/-10	2,452	Tencent	Proprietary	nan
69	61	o1-mini	1318	+3/-3	54,951	OpenAI	Proprietary	2023/10
70	61	o3-mini	1318	+3/-3	43,216	OpenAI	Proprietary	nan
71	62	Claude 3.7 Sonnet (thinking-32k)	1317	+4/-4	32,592	Anthropic	Proprietary	nan
72	62	Gemini-1.5-Pro-Exp-0827	1316	+4/-4	32,262	Google	Proprietary	2023/11
73	62	Gemini-1.5-Pro-Exp-0801	1315	+4/-4	25,988	Google	Proprietary	2023/11
74	60	Hunyuan-Turbo-0110	1313	+10/-10	2,510	Tencent	Proprietary	nan
75	62	Llama-3.3-Nemotron-Super-49B-v1	1309	+11/-11	2,371	Nvidia	Nvidia	nan
76	73	Grok-2-08-13	1305	+3/-3	67,084	xAI	Proprietary	2024/3
77	71	Gemma-3n-e4b-it	1305	+6/-6	13,182	Google	Gemma	nan
78	74	Yi-Lightning	1303	+4/-4	28,968	01 AI	Proprietary	nan
79	74	GPT-4o-2024-05-13	1302	+3/-3	117,747	OpenAI	Proprietary	2023/10
80	74	Claude 3.7 Sonnet	1301	+4/-4	37,645	Anthropic	Proprietary	nan
81	74	Qwen2.5-plus-1127	1300	+6/-6	10,715	Alibaba	Proprietary	nan
82	76	Claude 3.5 Sonnet (20241022)	1299	+2/-2	84,537	Anthropic	Proprietary	2024/4
83	74	Deepseek-v2.5-1210	1295	+7/-7	7,243	DeepSeek	DeepSeek	nan
84	80	Athene-v2-Chat-72B	1293	+4/-4	26,074	NexusFlow	NexusFlow	nan
85	76	Gemma-3-4B-it	1293	+9/-9	4,321	Google	Gemma	nan
86	81	Llama-4-Maverick-17B-128E-Instruct	1292	+4/-4	极24,544	Meta	Llama 4	nan
87	82	GLM-4-Plus	1291	+4/-4	27,788	Zhipu AI	Proprietary	nan
88	76	Hunyuan-Large-2025-02-10	1290	+9/-9	3,856	Tencent	Proprietary	nan
89	83	Gemini-1.5-Flash-002	1289	+3/-3	37,021	Google	Proprietary	nan
90	83	GPT-4o-mini-2024-07-18	1288	+3/-3	72,427	OpenAI	Proprietary	2023/10
91	82	GPT-4.1-nano-2025-04-14	1287	+7/-7	6,302	OpenAI	Proprietary	nan
92	83	Meta-Llama-3.1-405B-Instruct-bf16	1286	+3/-3	43,788	Meta	Llama 3.1 Community	2023/12
93	84	GPT-4o-2024-08-06	1285	+3/-3	47,973	OpenAI	Proprietary	2023/10
94	83	Llama-3.1-Nemotron-70B-Instruct	1284	+7/-7	7,577	Nvidia	Llama 3.1	2023/12
95	83	Qwen-Max-0919	1284	+5/-5	17,432	Alibaba	Qwen	nan
96	84	Gemini-1.5-Flash-Exp-0827	1284	+4/-4	25,409	Google	Proprietary	2023/11
97	85	Meta-Llama-3.1-405B-Instruct-fp8	1284	+3/-3	63,038	Meta	Llama 3.1 Community	2023/12
98	84	Yi-Lightning-lite	1283	+5/-5	17,067	01 AI	Proprietary	nan
99	87	Grok-2-Mini-08-13	1283	+3/-3	55,442	xAI	Proprietary	2024/3
100	87	Claude 3.5 Sonnet (20240620)	1283	+3/-3	86,159	Anthropic	Proprietary	2024/4
101	89	Gemini Advanced App (2024-05-14)	1280	+4/-4	52,144	Google	Proprietary	Online
102	95	Gemini-1.5-Pro-001	1276	+3/-3	82,435	Google	Proprietary	2023/11
103	91	Llama-4-Scout-17B-16E-Instruct	1276	+5/-5	14,153	Meta	Llama	nan
104	88	Hunyuan-Standard-2025-02-10	1276	+9/-9	4,014	Tencent	Proprietary	nan
105	99	Llama-3.3-70B-Instruct	1275	+3/-3	51,931	Meta	Llama-3.3	nan
106	100	GPT-4-Turbo-2024-04-09	1274	+3/-3	102,133	OpenAI	Proprietary	2023/12
107	99	Deepseek-v2.5	1274	+4/-4	26,344	DeepSeek	DeepSeek	nan
108	100	Gemini-1.5-Pro-Preview-0409	1273	+4/-4	55,569	Google	Proprietary	2023/11
109	101	Qwen2.5-72B-Instruct	1272	+3/-3	41,519	Alibaba	Qwen	2024/9
110	99	Hunyuan-Large-Vision	1270	+8/-8	6,093	Tencent	Proprietary	nan
111	102	Mistral-Large-2407	1269	+3/-3	48,217	Mistral	Mistral Research	2024/7
112	102	Mistral-Small-3.1-24B-Instruct-2503	1269	+5/-5	14,091	Mistral	Apache 2.0	nan
113	102	Mistral-Large-2411	1268	+4/-4	29,633	Mistral	MRL	nan
114	102	Athene-70B	1267	+5/-5	20,580	NexusFlow	CC-BY-NC-4.0	2024/7
115	107	GPT-4-1106-preview	1266	+3/-3	103,748	OpenAI	Proprietary	2023/4
116	107	GPT-4-0125-preview	1265	+3/-3	97,079	OpenAI	Proprietary	2023/12
117	109	Claude 3 Opus	1265	+2/-2	202,641	Anthropic	Proprietary	2023/8
118	109	Meta-Llama-3.1-70B-Instruct	1264	+3/-3	58,637	Meta	Llama 3.1 Community	2023/12
119	111	Amazon Nova Pro 1.0	1261	+4/-4	26,371	Amazon	Proprietary	nan
120	107	Llama-3.1-Tulu-3-70B	1259	+10/-10	3,010	Ai2	Llama 3.1	nan
121	114	Yi-Large-preview	1258	+4/-4	51,628	01 AI	Proprietary	データなし
122	119	Claude 3.5 Haiku (20241022)	1255	+3/-3	55,928	Anthropic	Propretary	nan
123	119	magistral-medium-2506	1252	+7/-7	8,534	Mistral	Proprietary	nan
124	119	Reka-Core-20240904	1251	+6/-6	7,948	Reka AI	Proprietary	nan
125	120	Reka-Core-20240722	1249	+6/-6	13,279	Reka AI	Proprietary	nan
126	123	Gemini-1.5-Flash-001	1243	+4/-4	65,661	Google	Proprietary	2023/11
127	123	Qwen-Plus-0828	1242	+5/-5	14,626	Alibaba	Proprietary	nan
128	123	Jamba-1.5-Large	1241	+6/-6	9,125	AI21 Labs	Jamba Open	2024/3
129	125	Deepseek-v2-API-0628	1239	+5/-5	19,508	DeepSeek AI	DeepSeek	nan
130	125	Mistral-Small-24B-Instruct-2501	1238	+5/-5	15,321	Mistral	Apache 2.0	nan
131	125	Deepseek-Coder-v2-0724	1236	+6/-6	11,725	DeepSeek	Proprietary	nan
132	126	Yi-Large	1235	+6/-6	16,624	01 AI	Proprietary	データなし
133	127	Gemma-2-27B-it	1235	+2/-2	79,538	Google	Gemma license	2024/6
134	126	Qwen2.5-Coder-32B-Instruct	1234	+7/-7	5,730	Alibaba	Apache 2.0	nan
135	126	Command R+ (08-2024)	1233	+6/-6	10,535	Cohere	CC-BY-NC-4.0	2024/8
136	127	Amazon Nova Lite 1.0	1233	+4/-4	20,646	Amazon	Proprietary	nan
137	126	Gemma-2-9B-it-SimPO	1232	+6/-6	10,548	Princeton	MIT	2024/极7
138	126	Llama-3.1-Nemotron-51B-Instruct	1231	+9/-9	3,889	Nvidia	Llama 3.1	2023/12
139	128	Gemini-1.5-Flash-8B-001	1230	+3/-3	37,697	Google	Proprietary	nan
140	127	GLM-4-0520	1230	+6/-6	10,221	Zhipu AI	Proprietary	データなし
141	130	Nemotron-4-340B-Instruct	1229	+4/-4	20,608	Nvidia	NVIDIA Open Model	2023/6
142	130	Aya-Expanse-32B	1228	+4/-4	28,768	Cohere	CC-BY-NC-4.0	nan
143	128	Gemini App (2024-01-24)	1227	+8/-8	11,833	Google	Proprietary	Online
144	132	Reka-Flash-20240904	1224	+6/-6	8,132	Reka AI	Proprietary	nan
145	135	Llama-3-70B-Instruct	1224	+3/-3	163,629	Meta	Llama 3 Community	2023/12
146	130	OLMo-2-0325-32B-Instruct	1223	+10/-10	3,460	Allen AI	Apache-2.0	nan
147	138	Claude 3 Sonnet	1222	+3/-3	113,067	Anthropic	Proprietary	2023/8
148	138	Phi-4	1222	+4/-4	25,213	Microsoft	MIT	nan
149	138	Gemini-1.5-Flash-8B-Exp-0827	1221	+4/-4	25,346	Google	Proprietary	2023/11
150	142	Reka-Core-20240501	1217	+3/-3	62,555	Reka AI	Proprietary	nan
151	142	Reka-Flash-20240722	1217	+6/-6	13,725	Reka AI	Proprietary	nan
152	144	Amazon Nova Micro 1.0	1214	+4/-4	20,654	Amazon	Proprietary	nan
153	149	Gemma-2-9B-it	1212	+3/-3	57,197	Google	Gemma license	2024/6
154	149	GPT-4-0314	1210	+4/-4	55,962	OpenAI	Proprietary	2021/9
155	151	Command R+ (04-2024)	1208	+3/-3	80,846	Cohere	CC-BY-NC-4.0	2024/3
156	144	Hunyuan-Standard-256K	1208	+11/-11	2,901	Tencent	Proprietary	nan
157	151	Qwen2-72B-Instruct	1208	+4/-4	38,872	Alibaba	Qianwen LICENSE	2024/6
158	156	Claude 3 Haiku	1200	+3/-3	122,309	Anthropic	Proprietary	2023/8
159	153	Llama-3.1-Tulu-3-8B	1199	+10/-10	3,074	Ai2	Llama 3.1	nan
160	156	Qwen-Max-0428	1199	+5/-5	25,696	Alibaba	Proprietary	nan
161	154	GLM-4-0116	1198	+9/-9	7,579	Zhipu AI	Proprietary	データなし
162	154	Ministral-8B-2410	1198	+8/-8	5,111	Mistral	MRL	nan
163	157	DeepSeek-Coder-V2-Instruct	1196	+6/-6	15,753	DeepSeek AI	DeepSeek License	2024/6
164	157	Command R (08-2024)	1195	+6/-6	10,851	Cohere	CC-BY-NC-4.0	2024/8
165	157	Aya-Expanse-8B	1193	+6/-6	10,391	Cohere	CC-BY-NC-4.0	nan
166	157	Jamba-1.5-Mini	1193	+6/-6	9,274	AI21 Labs	Jamba Open	2024/3
167	159	Meta-Llama-3.1-8B-Instruct	1193	+3/-3	52,578	Meta	Llama 3.1 Community	2023/12
168	159	GPT-4-0613	1191	+3/-3	91,614	OpenAI	Proprietary	2021/9
169	159	Reka-Flash-Preview-20240611	1188	+5/-5	20,427	Reka AI	Proprietary	nan
170	169	Mistral-Large-2402	1181	+4/-4	64,926	Mistral	Proprietary	nan
171	169	Qwen1.5-110B-Chat	1180	+5/-5	27,430	Alibaba	Qianwen LICENSE	2024/4
172	169	Claude-1	1178	+6/-6	21,149	Anthropic	Proprietary	nan
173	170	Yi-1.5-34B-Chat	1178	+4/-4	25,135	01 AI	Apache-2.0	2024/5
174	169	Reka-Flash-21B-online	1176	+7/-7	16,027	Reka AI	Proprietary	Online
175	170	Qwen1.5-72B-Chat	1171	+4/-4	40,658	Alibaba	Qianwen LICENSE	2024/2
176	172	Llama-3-8B-Instruct	1171	+3/-3	109,056	Meta	Llama 3 Community	2023/3
177	171	Mistral Medium	1170	+5/-5	35,556	Mistral	Proprietary	nan
178	171	Reka-Flash-21B	1170	+5/-5	25,803	Reka AI	Proprietary	2023/11
179	170	QwQ-32B-Preview	1170	+10/-10	3,410	Alibaba	Apache 2.0	nan
180	172	Command R (04-2024)	1169	+4/-4	56,398	Cohere	CC-BY-NC-4.0	2024/3
181	172	InternLM2.5-20B-chat	1168	+6/-6	10,599	InternLM	Other	2024/8
182	173	Mixtral-8x22b-Instruct-v0.1	1167	+4/-4	53,751	Mistral	Apache 2.0	2024/4
183	176	Gemma-2-2b-it	1163	+3/-3	48,892	Google	Gemma license	2024/7
184	177	Claude-2.0	1158	+8/-8	12,763	Anthropic	Proprietary	nan
185	175	Granite-3.1-8B-Instruct	1157	+10/-10	3,289	IBM	Apache 2.0	nan
186	181	Gemini-1.0-Pro-001	1154	+7/-7	18,800	Google	Proprietary	2023/4
187	181	Mistral-Next	1153	+8/-8	12,375	Mistral	Proprietary	nan
188	182	Zephyr-ORPO-141b-A35b-v0.1	1150	+10/-10	4,854	HuggingFace	Apache 2.0	2024/4
189	184	Claude-2.1	1146	+6/-6	37,699	Anthropic	Proprietary	nan
190	184	GPT-3.5-Turbo-061极3	1145	+5/-5	38,955	OpenAI	Proprietary	2021/9
191	184	Qwen1.5-32B-Chat	1144	+5/-5	22,765	Alibaba	Qianwen LICENSE	2024/2
192	185	Phi-3-Medium-4k-Instruct	1143	+4/-4	26,105	Microsoft	MIT	2023/10
193	187	Starling-LM-7B-beta	1138	+7/-7极	16,676	Nexusflow	Apache-2.0	2024/3
194	188	Mixtral-8x7B-Instruct-v0.1	1138	+3/-3	76,126	Mistral	Apache 2.0	2023/12
195	185	GPT-3.5-Turbo-0314	1137	+11/-11	5,640	OpenAI	Proprietary	2021/9
196	185	Gemini Pro	1137	+11/-11	6,557	Google	Proprietary	2023/4
197	187	Granite-3.1-2B-Instruct	1135	+10/-10	3,380	IBM	Apache 2.0	nan
198	188	Claude-Instant-1	1135	+7/-7	20,631	Anthropic	Proprietary	nan
199	188	Qwen1.5-14B-Chat	1135	+6/-6	18,687	Alibaba	Qianwen LICENSE	2024/2
200	188	Yi-34B-Chat	1134	+6/-6	15,917	01 AI	Yi License	2023/6
201	188	Yi-34B-Chat	1134	+6/-6	15,917	01 AI	Yi License	2023/6
202	193	GPT-3.5-Turbo-0125	1131	+4/-4	68,867	OpenAI	Proprietary	2021/9
203	193	Tulu-2-DPO-70B	1127	+9/-9	6,658	AllenAI/UW	AI2 ImpACT Low-risk	2023/11
204	195	DBRX-Instruct-Preview	1125	+5/-5	33,743	Databricks	DBRX LICENSE	2023/12
205	193	WizardLM-70B-v1.0	1125	+9/-9	8,383	Microsoft	Llama 2 Community	2023/8
206	199	Llama-2-70B-chat	1122	+5/-5	39,595	Meta	Llama 2 Community	2023/7
207	195	Nous-Hermes-2-Mixtral-8x7B-DPO	1119	+11/-11	3,836	NousResearch	Apache-2.0	2024/1
208	202	Meta-Llama-3.2-3B-Instruct	1117	+7/-7	8,390	Meta	Llama 3.2	2023/12
209	202	Phi-3-Small-8k-Instruct	1117	+5/-5	18,476	Microsoft	MIT	2023/10
210	202	Starling-LM-7B-alpha	1114	+7/-7	10,415	UC Berkeley	CC-BY-NC-4.0	2023/11
211	202	OpenChat-3.5-0106	1113	+7/-7	12,990	OpenChat	Apache-2.0	2024/1
212	203	Vicuna-33B	1112	+5/-5	22,936	LMSYS	Non-commercial	2023/8
213	202	DeepSeek-LLM-67B-Chat	1110	+11/-11	4,988	DeepSeek AI	DeepSeek License	2023/11
214	206	Snowflake Arctic Instruct	1108	+5/-5	34,173	Snowflake	Apache 2.0	2024/4
215	206	Granite-3.0-8B-Instruct	1107	+8/-8	7,002	IBM	Apache 2.0	nan
216	203	NV-Llama2-70B-SteerLM-Chat	1105	+12/-12	3,636	Nvidia	Llama 2 Community	2023/11
217	206	OpenChat-3.5	1103	+9/-9	8,106	OpenChat	Apache-2.0	2023/11
218	208	Gemma-1.1-7B-it	1102	+5/-5	25,070	Google	Gemma license	2024/2
219	208	GPT-3.5-Turbo-1106	1100	+8/-8	17,036	OpenAI	Proprietary	2021/9
220	208	OpenHermes-2.5-Mistral-7B	1100	+10/-10	5,088	NousResearch	Apache-2.0	2023/11
221	208	pplx-70B-online	1099	+10/-10	6,898	Perplexity AI	Proprietary	Online
222	212	Mistral-7B-Instruct-v0.2	1097	+6/-6	20,067	Mistral	Apache-2.0	2023/12
223	215	Llama-2-13b-chat	1092	+6/-6	19,722	Meta	Llama 2 Community	2023/7
224	212	SOLAR-10.7B-Instruct-v1.0	1091	+12/-12	4,286	Upstage AI	CC-BY-NC-4.0	2023/11
225	215	Granite-3.0-2B-Instruct	1090	+7/-7	7,191	IBM	Apache 2.0	nan
226	213	Qwen1.5-7B-Chat	1090	+9/-9	4,872	Alibaba	Qianwen LICENSE	2024/2
227	212	Dolphin-2.2.1-Mistral-7B	1088	+15/-15	1,714	Cognitive Computations	Apache-2.0	2023/10
228	216	Phi-3-Mini-4K-Instruct-June-24	1087	+6/-6	12,808	Microsoft	MIT	2023/10
229	217	WizardLM-13b-v1.2	1084	+9/-9	7,176	Microsoft	Llama 2 Community	2023/7
230	222	Phi-3-Mini-4k-Instruct	1081	+5/-5	21,097	Microsoft	MIT	2023/10
231	223	Zephyr-7B-beta	1076	+8/-8	11,321	HuggingFace	MIT	2023/10
232	222	MPT-30B-chat	1075	+12/-12	2,644	MosaicML	CC-BY-NC-SA-4.0	2023/6
233	225	CodeLlama-34B-instruct	1072	+8/-8	7,509	Meta	Llama 2 Community	2023/7
234	229	Meta-Llama-3.2-1B-Instruct	1067	+7/-7	8,523	Meta	Llama 3.2	2023/12
235	224	Zephyr-7B-alpha	1066	+15/-15	1,811	HuggingFace	MIT	2023/10
236	230	Vicuna-13B	1065	+6/-6	19,775	LMSYS	Llama 2 Community	2023/7
237	223	CodeLlama-70B-instruct	1065	+17/-17	1,192	Meta	Llama 2 Community	2024/1
238	229	pplx-7B-online	1064	+10/-10	6,339	Perplexity AI	Proprietary	Online
239	230	Gemma-7B-it	1063	+9/-9	9,176	Google	Gemma license	2024/2
240	228	SmolLM2-1.7B-Instruct	1062	+13/-13	2,375	HuggingFace	Apache 2.0	nan
241	226	falcon-180b-chat	1061	+17/-17	1,327	TII	Falcon-180B TII License	2023/9
242	230	Phi-3-Mini-128k-Instruct	1061	+6/-6	21,622	Microsoft	MIT	2023/10
243	230	Llama-2-7B-chat	1060	+6/-6	14,532	Meta	Llama 2 Community	2023/7
244	230	Guanaco-33B	1060	+12/-12	2,996	UW	Non-commercial	2023/5
245	230	Qwen-14B-Chat	1058	+10/-10	5,065	Alibaba	Qianwen LICENSE	2023/8
246	235	StripedHyena-Nous-7B	1044	+10/-10	5,276	Together AI	Apache 2.0	2023/12
247	243	Vicuna-7B	1038	+9/-9	7,017	LMSYS	Llama 2 Community	2023/7
248	240	OLMo-7B-instruct	1037	+10/-10	6,503	Allen AI	Apache-2.0	2024/2
249	245	PaLM-Chat-Bison-001	1034	+9/-9	8,713	Google	Proprietary	2021/6
250	245	Gemma-1.1-2b-it	1032	+7/-7	11,351	Google	Gemma license	2024/2
251	245	Mistral-7B-Instruct-v0.1	1031	+9/-9	9,142	Mistral	Apache 2.0	2023/9
252	251	Gemma-2B-it	1011	+11/-11	4,918	Google	Gemma license	2024/2
253	251	Qwen1.5-4B-Chat	1006	+9/-9	7,816	Alibaba	Qianwen LICENSE	2024/2
254	251	Koala-13B	996	+9/-9	7,020	UC Berkeley	Non-commercial	2023/4
255	253	ChatGLM3-6B	978	+11/-11	4,763	Tsinghua	Apache-2.0	2023/10
256	254	MPT-7B-Chat	963	+11/-11	3,997	MosaicML	CC-BY-NC-SA-4.0	2023/5
257	254	GPT4All-13B-Snoozy	962	+15/-15	1,788	Nomic AI	Non-commercial	2023/3
258	254	RWKV-4-Raven-14极B	956	+11/-11	4,920	RWKV	Apache 2.0	极2023/4
259	255	ChatGLM2-6B	947	+13/-13	2,713	Tsinghua	Apache-2.0	2023/6
260	255	Alpaca-13B	940	+11/-11	5,864	Stanford	Non-commercial	2023/3
261	258	ChatGLM-6B	925	+12/-12	4,983	Tsinghua	Non-commercial	2023/3
262	258	OpenAssistant-Pythia-12B	923	+10/-10	6,368	OpenAssistant	Apache 2.0	2023/4
263	260	FastChat-T5-3B	901	+12/-12	4,288	LMSYS	Apache 2.0	2023/4
264	263	StableLM-Tuned-Alpha-7B	873	+12/-12	3,336	Stability AI	CC-BY-NC-SA-4.0	2023/4
265	263	Dolly-V2-12B	857	+13/-13	3,480	Databricks	MIT	2023/4
266	264	LLaMA-13B	840	+16/-16	2,446	Meta	Non-commercial	2023/2

説明

• 順位(UB): Bradley-Terryモデルに基づくランキング。アリーナでのモデルの総合的なパフォーマンスを反映し、Eloスコアの上限推定を提供し、モデルの潜在的な競争力を理解するのに役立ちます。

• 順位(StyleCtrl): 会話スタイル制御後のランキング。モデルの応答スタイル（冗長さ、簡潔さなど）による嗜好の偏りを減らし、モデルの核心的な能力をより純粋に評価します。

• モデル名: 大規模言語モデル(LLM)の名前。関連リンクが埋め込まれており、クリックするとジャンプします。

• スコア: アリーナでのユーザー投票により獲得したElo評価スコア。Elo評価は相対的なランキングシステムで、スコアが高いほどモデルのパフォーマンスが優れていることを示します。このスコアは動的に変化し、現在の競争環境におけるモデルの相対的な実力を反映します。

• 信頼区間: モデルのElo評価スコアの95%信頼区間（例: +6/-6）。この区間が小さいほど、スコアが安定していて信頼性が高いことを示します。逆に区間が大きい場合は、データ量が不足しているか、モデルのパフォーマンスが変動しやすい可能性があります。スコアの正確性を定量化します。

• 投票数: アリーナでそのモデルが受けた総投票数。投票数が多いほど、通常、その評価の統計的信頼性が高くなります。

• プロバイダ: モデルを提供する組織または企業。

• ライセンス: モデルのライセンスタイプ（例: 専有(Proprietary)、Apache 2.0、MITなど）。

• ナレッジカットオフ: モデル訓練データの知識カットオフ日。データなしは情報が提供されていないか不明であることを示します。

データソースと更新頻度

このランキングのデータはfboulnois/llm-leaderboard-csvプロジェクトによって自動生成・提供されており、lmarena.aiからデータを取得・処理しています。このランキングはGitHub Actionsにより毎日自動更新されます。

免責事項

本レポートは参考情報としてご利用ください。ランキングデータは動的に変化し、特定の期間におけるChatbot Arena上でのユーザーの嗜好投票に基づいています。データの完全性と正確性は、上流のデータソースおよびfboulnois/llm-leaderboard-csvプロジェクトの更新と処理に依存します。異なるモデルは異なるライセンスを採用している場合があるため、ご利用の際は必ずモデルプロバイダの公式説明を参照してください。