Arena 网友盲测总榜

共 363 个模型，分数是原榜的；同一模型的不同档位只留最好的一档。

#	模型	机构	盲测分	全球名次	投票数
1	Gemini 4 Carbon	google	1528 ±9	1	4,403
2	Gemini 4 Argon	google	1520 ±6	2	11,889
3	Claude Opus 5.5	anthropic	1509 ±7	3	11,415
4	Muse Wonder	meta	1507 ±7	4	7,749
5	Claude Fable 5.1	anthropic	1504 ±6	5	17,178
6	Claude Fable 5	anthropic	1504 ±4	6	45,341
7	Claude Opus 4.6	anthropic	1503 ±3	7	84,512
8	Claude Opus 4.7	anthropic	1501 ±4	8	65,110
9	Gemini 3.8 Flash	google	1496 ±4	10	47,803
10	Muse Spark 1.3	meta	1495 ±5	11	22,134
11	Muse Spark 1.2	meta	1492 ±6	13	11,724
12	Claude Opus 5	anthropic	1492 ±4	14	73,167
13	Muse Spark 1.1	meta	1490 ±4	16	46,354
14	Gemini 3.7 Flash	google	1490 ±5	17	32,289
15	Muse Spark	meta	1489 ±6	18	14,140
16	Kimi K3	moonshot	1488 ±4	19	36,780
17	Gemini 3.1 Pro Preview	google	1487 ±3	20	132,547
18	Gemini 3 Pro	google	1485 ±4	21	42,057
19	GPT-5.6 Sol	openai	1485 ±4	22	43,796
20	Gemini 3.6 Flash	google	1483 ±4	23	36,028
21	Qwen3.8-Max · 千问	alibaba	1482 ±5	24	31,915
22	Claude Opus 4.8	anthropic	1482 ±4	25	63,994
23	GPT-5.5	openai	1482 ±4	26	69,377
24	GPT-6.1 Sol	openai	1481 ±7	27	9,367
25	GLM-5.3	zai	1481 ±5	28	26,606
26	MiMo-V2.6-Pro	xiaomi	1480 ±9	29	4,037
27	Gemini 3.5 Flash	google	1478 ±4	30	48,376
28	Claude Sonnet 5.5	anthropic	1477 ±7	31	10,033
29	GLM-5.3-Flash	zai	1477 ±5	33	41,642
30	DeepSeek-V4.1-Flash-20260910	deepseek	1476 ±6	35	13,657
31	GPT-5.2 Chat	openai	1476 ±4	36	35,974
32	GPT-6 Astra	openai	1476 ±6	37	15,038
33	GLM-5.2	zai	1475 ±4	38	53,702
34	GPT-5.4	openai	1475 ±4	39	64,748
35	Grok 4.20 Beta 1	xai	1475 ±5	40	27,889
36	Qwen3.7-Max-Preview · 千问	alibaba	1474 ±10	42	3,932
37	Claude Opus 4.5	anthropic	1474 ±4	43	37,484
38	GPT-5.5 Instant	openai	1473 ±5	44	27,128
39	Gemini 3 Flash	google	1473 ±4	45	31,301
40	Claude Sonnet 4.6	anthropic	1472 ±4	46	70,916
41	Grok 4.20 Beta	xai	1472 ±4	47	66,422
42	Grok 4.20 Multi-Agent Beta	xai	1471 ±4	48	64,917
43	GPT-5.6 Terra	openai	1467 ±4	50	48,166
44	ERNIE 5.1 · 文心	baidu	1467 ±5	51	39,469
45	MiMo-V2.5-Pro	xiaomi	1467 ±4	52	77,728
46	Grok 4.5	xai	1465 ±4	53	39,801
47	GLM-5.1	zai	1465 ±4	54	66,771
48	DeepSeek-V4-Pro	deepseek	1465 ±7	55	9,787
49	Grok 4.1	xai	1465 ±3	56	66,470
50	Qwen3.5-Max-Preview · 千问	alibaba	1465 ±5	58	22,582
51	Claude Sonnet 5	anthropic	1462 ±4	59	52,952
52	Kimi K2.6	moonshot	1461 ±4	60	39,978
53	Qwen3.6-Max-Preview · 千问	alibaba	1460 ±8	61	5,419
54	GPT-6 Sol	openai	1458 ±6	64	15,807
55	GLM-5	zai	1457 ±4	66	29,255
56	Claude Sonnet 4.5	anthropic	1456 ±3	67	83,892
57	Dola Seed 2.0 Pro · 豆包	bytedance	1456 ±3	68	79,094
58	GPT-5.1	openai	1456 ±4	69	41,639
59	Qwen3.7-Plus · 千问	alibaba	1456 ±4	70	42,011
60	Gemini 3.5 Flash-Lite	google	1456 ±4	71	35,740
61	Grok 4.6	xai	1455 ±5	72	33,824
62	DeepSeek-V4-Pro Preview	deepseek	1455 ±4	74	54,870
63	Hy3 · 元宝	tencent	1454 ±6	75	13,840
64	GPT-5.6 Luna	openai	1453 ±4	76	41,317
65	Gemma 4 31B · Gemini	google	1452 ±7	77	6,138
66	GPT-5.3 Chat	openai	1450 ±4	78	34,316
67	Claude Opus 4.1	anthropic	1450 ±3	79	49,535
68	Kimi K2.5	moonshot	1450 ±3	80	74,682
69	ERNIE-5.0-Preview · 文心	baidu	1448 ±6	81	9,802
70	Step 5 Preview	stepfun	1448 ±6	82	11,355
71	MiMo-V2-Pro	xiaomi	1448 ±5	84	25,421
72	MiMo-V2.6-Flash	xiaomi	1448 ±6	85	15,312
73	GPT-5.4 mini	openai	1447 ±4	86	63,432
74	ERNIE 5.0 · 文心	baidu	1446 ±4	87	36,475
75	Gemini 2.5 Pro	google	1445 ±2	88	125,307
76	GPT-4.5 Preview	openai	1445 ±6	89	14,547
77	Qwen3.6-Plus · 千问	alibaba	1443 ±4	90	48,126
78	ChatGPT-4o Latest	openai	1443 ±3	91	81,893
79	Inkling	thinky	1442 ±4	92	45,472
80	Grok 4.3	xai	1442 ±4	93	71,287
81	Qwen3.5-397B-A17B · 千问	alibaba	1441 ±3	94	97,675
82	GLM-4.7	zai	1441 ±6	95	12,239
83	Grok 4.7	xai	1441 ±6	96	16,442
84	GPT-6 Luna	openai	1441 ±4	97	86,478
85	MiniMax-M3	minimax	1439 ±4	98	68,611
86	DeepSeek-V4-Flash Preview	deepseek	1439 ±4	99	51,763
87	Gemma 4 26B A4B · Gemini	google	1438 ±7	101	6,090
88	GPT-5.2	openai	1437 ±4	102	49,628
89	LongCat-Flash-Chat Exp	meituan	1436 ±5	103	29,244
90	DeepSeek-V4-Flash	deepseek	1436 ±4	104	51,944
91	Beam 1.0 501B A23B	reflection	1436 ±9	105	4,233
92	Qwen3.8-27B · 千问	alibaba	1436 ±5	106	27,779
93	GPT-5	openai	1435 ±5	108	31,718
94	Qwen3-Max-Preview · 千问	alibaba	1434 ±5	109	27,368
95	GLM-5V-Turbo	zai	1434 ±6	110	10,476
96	MiMo-V2.5	xiaomi	1434 ±4	111	47,327
97	Gemini 3.1 Flash-Lite Preview	google	1433 ±4	112	64,196
98	Mistral Large 4	mistral	1432 ±6	113	12,066
99	o3 · ChatGPT	openai	1432 ±4	114	58,901
100	MiMo-V2-Omni	xiaomi	1431 ±6	115	20,615
101	Grok 4.1 Fast	xai	1430 ±3	117	57,500
102	Kimi K2 Thinking Turbo	moonshot	1430 ±3	118	63,310
103	amazon-nova-experimental-chat	amazon	1427 ±10	119	3,523
104	NVIDIA-Nemotron-3-Ultra-550B-A55B	nvidia	1427 ±7	120	11,563
105	GPT-5 Chat	openai	1426 ±4	121	31,187
106	Mistral Medium 3.5	mistral	1426 ±6	122	11,791
107	Claude Opus 4	anthropic	1426 ±4	123	35,904
108	DeepSeek-V3.2-Exp	deepseek	1425 ±7	124	8,991
109	Muse Glimmer	meta	1425 ±10	125	3,884
110	Claude Haiku 5.5	anthropic	1425 ±7	126	7,890
111	DeepSeek-V3.2	deepseek	1425 ±3	127	48,095
112	GLM-4.6	zai	1424 ±4	128	35,810
113	Qwen3-Max · 千问	alibaba	1423 ±6	129	9,026
114	Qwen3-235B-A22B-Instruct · 千问	alibaba	1422 ±3	131	98,081
115	DeepSeek-R1	deepseek	1421 ±6	133	18,165
116	Kimi K2 Preview	moonshot	1418 ±7	134	11,592
117	Grok 4 Fast	xai	1418 ±8	135	6,336
118	DeepSeek-V3.1	deepseek	1417 ±6	138	14,606
119	Qwen3.5-122B-A10B · 千问	alibaba	1416 ±4	139	29,712
120	DeepSeek-V3.1-Terminus	deepseek	1416 ±10	140	3,379
121	MiniMax-M2.7	minimax	1415 ±3	142	89,197
122	GPT-4.1	openai	1415 ±4	144	50,298
123	Claude Haiku 4.5	anthropic	1415 ±2	145	154,063
124	Mistral Large 3	mistral	1414 ±3	146	88,325
125	Qwen3-VL-235B-A22B-Instruct · 千问	alibaba	1413 ±7	148	11,314
126	NVIDIA-Nemotron-3.5-Super-VL-120B-A12B	nvidia	1412 ±9	150	4,178
127	Grok 3 Preview	xai	1411 ±4	151	32,496
128	Grok 4	xai	1411 ±4	152	39,678
129	GLM-4.5	zai	1411 ±5	153	23,797
130	Gemini 2.5 Flash	google	1409 ±2	154	125,457
131	Qwen3.5-27B · 千问	alibaba	1409 ±4	155	28,647
132	Hy3 preview · 元宝	tencent	1409 ±7	156	6,869
133	Mistral Medium 3.1	mistral	1408 ±3	157	95,269
134	Inkling-Small	thinky	1406 ±5	158	36,999
135	Gemini 2.5 Flash Preview	google	1403 ±4	160	32,996
136	o1 · ChatGPT	openai	1402 ±4	161	27,807
137	Qwen3-235B-A22B · 千问	alibaba	1402 ±5	162	37,472
138	Claude Sonnet 4	anthropic	1402 ±4	163	34,128
139	GPT-5.4 nano	openai	1401 ±4	164	62,314
140	LongCat-Flash-Chat	meituan	1401 ±6	165	11,167
141	Qwen3-235B-A22B-Thinking · 千问	alibaba	1400 ±7	166	8,822
142	Qwen3-Next-80B-A3B-Instruct · 千问	alibaba	1399 ±5	167	22,753
143	Qwen3.5-Flash · 千问	alibaba	1396 ±4	169	61,776
144	DeepSeek-V3	deepseek	1395 ±4	170	44,977
145	Qwen3-VL-235B-A22B-Thinking · 千问	alibaba	1395 ±7	171	7,862
146	Hunyuan Vision 1.5 Thinking · 元宝	tencent	1394 ±12	172	2,169
147	Qwen3.5-35B-A3B · 千问	alibaba	1394 ±4	173	30,427
148	Step-3.5-Flash	stepfun	1393 ±4	174	60,305
149	MiMo-V2-Flash	xiaomi	1392 ±3	176	48,060
150	o4-mini · ChatGPT	openai	1391 ±4	177	44,843
151	MiniMax-M2.5	minimax	1391 ±4	179	42,783
152	GPT-5 mini	openai	1390 ±5	180	26,793
153	o1-preview · ChatGPT	openai	1389 ±5	181	31,122
154	Claude 3.7 Sonnet	anthropic	1388 ±4	182	38,156
155	Solar Pro 4	upstage	1388 ±5	183	26,159
156	Qwen3-Coder-480B-A35B-Instruct · 千问	alibaba	1388 ±5	184	25,190
157	Hunyuan T1 · 元宝	tencent	1387 ±9	185	4,595
158	Mistral Medium 3	mistral	1387 ±5	186	32,777
159	MiniMax-M2.1 Preview	minimax	1385 ±5	188	17,222
160	Hunyuan TurboS · 元宝	tencent	1383 ±6	189	10,623
161	GPT-4.1 mini	openai	1383 ±4	190	38,696
162	Qwen3-30B-A3B-Instruct · 千问	alibaba	1382 ±5	191	23,312
163	Gemini 2.5 Flash-Lite Preview	google	1379 ±3	192	47,418
164	Trinity Large Preview	—	1379 ±4	193	31,233
165	GLM-4.6V	zai	1378 ±11	194	2,833
166	Claude 3.5 Sonnet	anthropic	1374 ±3	197	87,771
167	Qwen2.5-Max · 千问	alibaba	1374 ±4	198	32,487
168	GLM-4.5-Air	zai	1373 ±4	199	30,652
169	Qwen3-Next-80B-A3B-Thinking · 千问	alibaba	1369 ±6	201	13,476
170	Trinity Large Thinking	—	1367 ±5	202	30,290
171	Gemma 3 27B IT · Gemini	google	1365 ±4	203	46,670
172	GLM-4.7-Flash	zai	1365 ±6	204	12,008
173	MiniMax-M1	minimax	1364 ±4	205	34,293
174	Grok 3 Mini	xai	1364 ±5	207	15,595
175	o3-mini · ChatGPT	openai	1364 ±5	208	18,589
176	Gemini 2.0 Flash	google	1360 ±4	209	43,443
177	NVIDIA-Nemotron-3-Super-120B-A12B	nvidia	1360 ±7	210	7,870
178	Grok 3 Mini Beta	xai	1358 ±5	212	20,930
179	Mistral Small 3.2	mistral	1356 ±5	213	17,462
180	INTELLECT-3	—	1356 ±8	214	5,435
181	Command A	cohere	1354 ±3	215	55,763
182	Gemini 2.0 Flash-Lite Preview	google	1354 ±4	216	24,955
183	NVIDIA-Nemotron-3.5-Lightning-30B-A3B	nvidia	1353 ±5	217	24,230
184	GLM-4.5V	zai	1352 ±8	218	4,817
185	gpt-oss-120b	openai	1352 ±4	219	30,200
186	Gemini 1.5 Pro	google	1351 ±3	220	55,606
187	Step-3	stepfun	1349 ±8	221	6,291
188	Llama 3.1 Nemotron Ultra 253B v1	nvidia	1348 ±12	225	2,549
189	Qwen3-32B · 千问	alibaba	1347 ±9	226	3,926
190	GPT-4o	openai	1346 ±3	228	112,881
191	Qwen-Plus · 千问	alibaba	1346 ±8	229	5,819
192	MiniMax-M2	minimax	1344 ±8	230	6,850
193	Ling-flash-2.0	ant-group	1343 ±7	232	6,952
194	Llama 3.3 Nemotron Super 49B v1.5	nvidia	1343 ±10	233	3,327
195	Mercury 2	inception-ai	1343 ±11	234	3,191
196	GLM-4-Plus	zai	1343 ±8	235	5,760
197	Gemma 3 12B IT · Gemini	google	1342 ±10	236	3,829
198	Hunyuan Turbo · 元宝	tencent	1341 ±12	237	2,290
199	Granite 4.2 30B	ibm	1340 ±10	238	3,389
200	GPT-5 nano	openai	1338 ±7	239	8,134
201	o1-mini · ChatGPT	openai	1337 ±4	240	51,981
202	Gemini Advanced	google	1336 ±5	241	50,148
203	Grok 2	xai	1336 ±4	242	63,498
204	QwQ-32B · 千问	alibaba	1336 ±4	244	25,177
205	Llama 3.1 405B Instruct	meta	1335 ±4	245	41,375
206	Amazon Nova 2 Lite	amazon	1335 ±6	246	12,415
207	Step-2 16K Exp	stepfun	1334 ±9	248	4,833
208	Olmo 3.1 32B Instruct	allenai	1329 ±6	249	11,805
209	Yi Lightning	—	1328 ±5	250	27,332
210	Llama 3.3 Nemotron Super 49B v1	nvidia	1328 ±12	251	2,218
211	Llama 4 Maverick 17B 128E Instruct	meta	1327 ±4	252	39,439
212	Qwen3-30B-A3B · 千问	alibaba	1327 ±5	253	26,156
213	Hunyuan Large · 元宝	tencent	1326 ±10	254	3,738
214	Claude 3.5 Haiku	anthropic	1325 ±3	255	69,056
215	GPT-4 Turbo	openai	1325 ±4	256	98,114
216	DeepSeek-V2.5	deepseek	1323 ±8	258	6,795
217	Ring-flash-2.0	ant-group	1322 ±7	259	6,987
218	GPT-4.1 nano	openai	1322 ±8	260	6,103
219	Claude 3 Opus	anthropic	1322 ±3	261	194,909
220	Llama 4 Scout 17B 16E Instruct	meta	1321 ±5	262	29,823
221	Molmo 2 8B	allenai	1321 ±21	263	822
222	Step-1o Turbo	stepfun	1320 ±7	264	8,906
223	Qwen-Max · 千问	alibaba	1318 ±6	266	16,478
224	Llama 3.3 70B Instruct	meta	1318 ±4	267	54,493
225	GPT-4o mini	openai	1318 ±4	268	68,705
226	Gemma 3n E4B IT · Gemini	google	1318 ±5	269	22,136
227	gpt-oss-20b	openai	1317 ±6	270	10,416
228	Qwen2.5-Plus · 千问	alibaba	1315 ±6	271	10,187
229	Mistral Large 2	mistral	1314 ±4	272	45,459
230	Athene V2 Chat	—	1314 ±5	273	24,739
231	NVIDIA-Nemotron-3-Nano-30B-A3B	nvidia	1314 ±5	274	15,761
232	GPT-4 Turbo Preview	openai	1313 ±4	275	93,439
233	Hunyuan Standard · 元宝	tencent	1311 ±10	277	3,904
234	Gemini 1.5 Flash	google	1309 ±4	278	34,902
235	Grok 2 mini	xai	1308 ±4	279	52,567
236	Athene 70B	—	1307 ±6	281	19,621
237	Olmo 3 32B Think	allenai	1306 ±8	282	5,827
238	Mercury	inception-ai	1306 ±14	283	1,966
239	Mistral Large	mistral	1306 ±4	284	28,073
240	Magistral Medium	mistral	1305 ±6	285	11,513
241	Granite 4.1 8B	ibm	1304 ±10	286	4,177
242	Gemma 3 4B IT · Gemini	google	1303 ±9	287	4,171
243	Qwen2.5-72B-Instruct · 千问	alibaba	1303 ±4	288	39,406
244	Mistral Small 3.1 24B Instruct	mistral	1303 ±5	289	32,722
245	Llama 3.1 Nemotron 70B Instruct	nvidia	1299 ±8	290	7,140
246	Hunyuan Large Vision · 元宝	tencent	1294 ±9	291	5,279
247	Llama 3.1 70B Instruct	meta	1293 ±4	292	55,240
248	Amazon Nova Pro 1.0	amazon	1290 ±5	293	24,745
249	Granite 4.2 3B	ibm	1290 ±12	294	3,249
250	Granite 4.2 8B	ibm	1290 ±11	295	3,251
251	Gemma 2 27B IT · Gemini	google	1289 ±3	296	75,754
252	Jamba 1.5 Large	—	1289 ±7	297	8,662
253	Reka Core	—	1288 ±7	298	7,312
254	GPT-4	openai	1288 ±5	299	54,173
255	Granite-4.0-H-Small	ibm	1287 ±8	300	5,630
256	Llama 3.1 Nemotron 51B Instruct	nvidia	1287 ±10	301	3,749
257	Olmo 3.1 32B Think	allenai	1286 ±7	303	8,299
258	Llama 3.1 Tulu 3 70B	allenai	1286 ±10	304	2,846
259	Claude 3 Sonnet	anthropic	1281 ±4	305	109,284
260	Gemma 2 9B IT SimPO	—	1280 ±7	306	10,072
261	Nemotron-4 340B Instruct	nvidia	1277 ±5	307	19,659
262	Llama 3 70B Instruct	meta	1277 ±4	308	156,876
263	Command R+	cohere	1276 ±7	309	9,866
264	Mistral Small 24B Instruct	mistral	1274 ±6	311	14,681
265	GLM-4	zai	1273 ±7	312	9,788
266	Reka Flash	—	1272 ±7	313	7,536
267	Qwen2.5-Coder-32B-Instruct · 千问	alibaba	1271 ±8	314	5,432
268	Aya Expanse 32B	cohere	1267 ±5	315	27,124
269	Gemma 2 9B IT · Gemini	google	1267 ±4	316	54,611
270	DeepSeek-Coder-V2	deepseek	1265 ±6	317	15,147
271	Qwen2-72B-Instruct · 千问	alibaba	1262 ±5	318	37,325
272	Claude 3 Haiku	anthropic	1262 ±4	320	117,701
273	Amazon Nova Lite 1.0	amazon	1260 ±5	321	19,372
274	Gemini 1.5 Flash-8B	google	1259 ±4	322	35,558
275	Phi-4	microsoft	1256 ±5	323	24,126
276	OLMo 2 32B Instruct	allenai	1252 ±11	324	3,334
277	Amazon Nova Micro 1.0	amazon	1241 ±5	327	19,364
278	Jamba 1.5 Mini	—	1240 ±7	328	8,858
279	Ministral 8B	mistral	1238 ±9	329	4,781
280	Gemini Pro	google	1237 ±7	330	18,354
281	Qwen1.5-110B-Chat · 千问	alibaba	1234 ±6	331	26,195
282	Reka Flash 21B	—	1234 ±7	332	15,450
283	Qwen1.5-72B-Chat · 千问	alibaba	1234 ±5	333	39,302
284	Hunyuan Standard 256K · 元宝	tencent	1233 ±12	334	2,728
285	Mixtral 8x22B Instruct v0.1	mistral	1230 ±5	335	51,416
286	GPT-3.5 Turbo	openai	1226 ±5	338	66,207
287	Llama 3 8B Instruct	meta	1224 ±4	339	104,642
288	Aya Expanse 8B	cohere	1223 ±7	341	9,818
289	Mistral Medium	mistral	1223 ±5	342	34,550
290	Llama 3.1 Tulu 3 8B	allenai	1221 ±11	343	2,896
291	Zephyr ORPO 141B A35B v0.1	—	1213 ±11	344	4,652
292	Yi 1.5 34B Chat	—	1213 ±5	345	24,146
293	Llama 3.1 8B Instruct	meta	1211 ±4	346	49,605
294	Granite 3.1 8B Instruct	ibm	1209 ±11	347	3,090
295	Qwen1.5-32B-Chat · 千问	alibaba	1204 ±6	349	21,741
296	Gemma 2 2B IT · Gemini	google	1200 ±4	350	46,616
297	Phi-3 Medium 4K Instruct	microsoft	1198 ±5	351	25,055
298	Mixtral 8x7B Instruct v0.1	mistral	1197 ±4	352	73,503
299	DBRX Instruct Preview	—	1196 ±6	353	32,191
300	Qwen1.5-14B-Chat · 千问	alibaba	1191 ±7	354	17,839
301	InternLM2.5 20B Chat	—	1191 ±7	355	9,901
302	DeepSeek LLM 67B Chat	deepseek	1185 ±11	356	4,932
303	WizardLM 70B	microsoft	1185 ±9	357	8,214
304	Yi 34B Chat	—	1184 ±7	358	15,483
305	Granite 3.0 8B Instruct	ibm	1183 ±9	359	6,638
306	OpenChat 3.5	—	1183 ±10	360	7,968
307	Gemma 1.1 7B IT · Gemini	google	1183 ±6	362	23,893
308	Snowflake Arctic Instruct	—	1180 ±6	363	32,832
309	Granite 3.1 2B Instruct	ibm	1179 ±11	364	3,188
310	Tulu 2 DPO 70B	—	1178 ±10	365	6,535
311	OpenHermes 2.5 Mistral 7B	—	1176 ±10	366	5,006
312	Vicuna 33B	—	1173 ±6	367	22,479
313	Phi-3 Small 8K Instruct	microsoft	1171 ±6	368	17,766
314	Starling LM 7B Beta	—	1171 ±7	369	16,056
315	Llama 2 70B Chat	meta	1171 ±5	370	38,492
316	Starling LM 7B Alpha	—	1167 ±8	371	10,224
317	Llama 3.2 3B Instruct	meta	1167 ±8	372	7,936
318	Nous Hermes 2 Mixtral 8x7B DPO	—	1164 ±12	373	3,777
319	Granite 3.0 2B Instruct	ibm	1157 ±8	374	6,837
320	Llama 2 70B SteerLM Chat	nvidia	1155 ±13	375	3,585
321	QwQ-32B-Preview · 千问	alibaba	1154 ±11	376	3,231
322	SOLAR 10.7B Instruct v1.0	upstage	1152 ±13	377	4,155
323	Dolphin 2.2.1 Mistral 7B	—	1152 ±15	378	1,679
324	MPT 30B Chat	—	1151 ±12	379	2,572
325	WizardLM 13B	microsoft	1150 ±9	380	7,044
326	Mistral 7B Instruct v0.2	mistral	1149 ±7	381	19,402
327	Falcon 180B Chat	—	1148 ±17	382	1,295
328	Qwen1.5-7B-Chat · 千问	alibaba	1144 ±10	383	4,737
329	Phi-3 Mini 4K Instruct	microsoft	1143 ±6	384	12,297
330	Vicuna 13B	—	1142 ±7	385	19,367
331	Llama 2 13B Chat	meta	1141 ±7	386	19,174
332	Qwen-14B-Chat · 千问	alibaba	1139 ±11	387	4,964
333	PaLM 2 · Gemini	google	1139 ±9	388	8,554
334	Gemma 7B IT · Gemini	google	1138 ±9	389	8,925
335	Code Llama 34B Instruct	meta	1137 ±9	390	7,366
336	Zephyr 7B Beta	—	1131 ±9	391	11,118
337	Phi-3 Mini 128K Instruct	microsoft	1130 ±7	392	20,685
338	Guanaco 33B	—	1128 ±12	394	2,921
339	Zephyr 7B Alpha	—	1127 ±16	395	1,785
340	StripedHyena Nous 7B	—	1122 ±11	396	5,182
341	Code Llama 70B Instruct	meta	1119 ±18	397	1,143
342	Gemma 1.1 2B IT · Gemini	google	1117 ±8	398	10,854
343	Vicuna 7B	—	1115 ±9	399	6,923
344	SmolLM2 1.7B Instruct	—	1115 ±14	400	2,199
345	Llama 3.2 1B Instruct	meta	1111 ±8	401	8,045
346	Mistral 7B Instruct	mistral	1110 ±9	402	8,977
347	Llama 2 7B Chat	meta	1108 ±7	403	14,148
348	Gemma 2B IT · Gemini	google	1094 ±11	404	4,780
349	Qwen1.5-4B-Chat · 千问	alibaba	1091 ±9	405	7,597
350	OLMo 7B Instruct	allenai	1074 ±11	406	6,328
351	Koala 13B	—	1071 ±10	407	6,965
352	Alpaca 13B	—	1070 ±11	408	5,745
353	GPT4All 13B Snoozy	—	1068 ±15	409	1,743
354	MPT 7B Chat	—	1064 ±12	410	3,924
355	ChatGLM3 6B	—	1057 ±12	411	4,658
356	RWKV-4 Raven 14B	—	1042 ±11	412	4,845
357	ChatGLM2 6B	—	1025 ±14	413	2,658
358	OpenAssistant Pythia 12B	—	1024 ±11	414	6,310
359	ChatGLM 6B	—	996 ±12	415	4,914
360	FastChat-T5 3B	—	993 ±12	416	4,203
361	Dolly 2.0 12B	—	983 ±13	417	3,412
362	LLaMA 13B	meta	975 ±16	418	2,391
363	StableLM Tuned Alpha 7B	—	953 ±13	419	3,287

数据：Arena 网友盲测（2026-10-28 发布）、Epoch AI 基准成绩（2026-10-30 取得），均为 CC BY 4.0，详见关于。产品信息（国内能否直接用、免费版、官网）整理于 2026-10-27，以官网为准。