2026 年 11 月 AI 模型榜单

分类显示原始分；总榜以 1000－来源名次汇总分类序位，不表示跨任务能力分。11 月评分口径改变，不计算虚假涨跌。核验日：2026-11-03；Arena 数据日：2026-10-27。

已核验 304 条分类记录。历史未核验记录不计分；样本来自不同子榜，不代表独立用户总数。

text

来源名次	模型（含配置）	原始分	样本数
1	Gemini 4 Carbon (High)	1528±9 Preliminary	4403
2	Gemini 4 Argon (High)	1520±6	11889
3	Claude Opus 5.5 (High)	1509±7	11415
4	Muse Wonder (Max)	1507±7	7749
5	Claude Fable 5.1 (Max)	1504±6	17178
6	Claude Fable 5 (High)	1504±4	45341
7	Claude Opus 4.6 (High)	1503±3	84512
8	Claude Opus 4.7 (High)	1501±4	65110
9	Claude Opus 4.6	1496±3	89324
10	Gemini 3.8 Flash (High)	1496±4 Preliminary	47803
11	Muse Spark 1.3 (Max)	1495±5	22134
12	Claude Opus 4.7	1494±4	66140
13	Muse Spark 1.2 (xHigh)	1492±6	11724
14	Claude Opus 5 (High)	1492±4	73167
15	Claude Opus 5 (Max)	1491±5	35531
16	Muse Spark 1.1	1490±4	46354
17	Gemini 3.7 Flash (High)	1490±5 Preliminary	32289
18	Muse Spark	1489±6	14140
19	Kimi K3 (Max)	1488±4	36780
20	Gemini 3.1 Pro Preview	1487±3	132547
21	Gemini 3 Pro	1485±4	42057
22	GPT-5.6 Sol (xHigh)	1485±4	43796
23	Gemini 3.6 Flash (High)	1483±4	36028
24	Qwen3.8-Max	1482±5	31915
25	Claude Opus 4.8 (High)	1482±4	63994
26	GPT-5.5 (High)	1482±4	69377
27	GPT-6.1 Sol (Max)	1481±7	9367
28	GLM-5.3 (Max)	1481±5	26606
29	MiMo-V2.6-Pro	1480±9	4037
30	Gemini 3.5 Flash (High)	1478±4	48376
31	Claude Sonnet 5.5 (xHigh)	1477±7	10033
32	GPT-5.5	1477±4	70949
33	GLM-5.3-Flash	1477±5	41642
34	Gemini 3.5 Flash (Medium)	1476±4	47196
35	DeepSeek-V4.1-Flash-20260910 (Max)	1476±6	13657
36	GPT-5.2 Chat (20260210)	1476±4	35974
37	GPT-6 Astra (Max)	1476±6	15038
38	GLM-5.2 (Max)	1475±4	53702
39	GPT-5.4 (High)	1475±4	64748
40	Grok 4.20 Beta 1	1475±5	27889
41	Claude Opus 4.8	1474±4	65121
42	Qwen3.7-Max-Preview	1474±10	3932
43	Claude Opus 4.5 (High, 32K, 20251101)	1474±4	37484
44	GPT-5.5 Instant	1473±5	27128
45	Gemini 3 Flash	1473±4	31301
46	Claude Sonnet 4.6	1472±4	70916
47	Grok 4.20 Beta (Reasoning, 0309)	1472±4	66422
48	Grok 4.20 Multi-Agent Beta (0309)	1471±4	64917
49	Claude Opus 4.5 (20251101)	1470±3	72825
50	GPT-5.6 Terra (xHigh)	1467±4	48166
51	ERNIE 5.1	1467±5	39469
52	MiMo-V2.5-Pro	1467±4	77728
53	Grok 4.5	1465±4	39801
54	GLM-5.1	1465±4	66771
55	DeepSeek-V4-Pro (High, 20260813)	1465±7	9787
56	Grok 4.1 (Thinking)	1465±3	66470
57	GPT-5.4	1465±4	67856
58	Qwen3.5-Max-Preview	1465±5	22582
59	Claude Sonnet 5 (High)	1462±4	52952
60	Kimi K2.6	1461±4	39978
61	Qwen3.6-Max-Preview	1460±8	5419
62	Grok 4.1	1459±3	68599
63	Gemini 3 Flash (Minimal)	1458±3	90997
64	GPT-6 Sol (Max)	1458±6	15807
65	DeepSeek-V4-Pro	1458±4	57572
66	GLM-5	1457±4	29255
67	Claude Sonnet 4.5 (High, 32K, 20250929)	1456±3	83892
68	Dola Seed 2.0 Pro	1456±3	79094
69	GPT-5.1 (High)	1456±4	41639
70	Qwen3.7-Plus	1456±4	42011
71	Gemini 3.5 Flash-Lite	1456±4	35740
72	Grok 4.6 (High)	1455±5	33824
73	Claude Sonnet 4.5 (20250929)	1455±3	82621
74	DeepSeek-V4-Pro Preview (High)	1455±4	54870
75	Hy3	1454±6	13840
76	GPT-5.6 Luna (xHigh)	1453±4	41317
77	Gemma 4 31B	1452±7	6138
78	GPT-5.3 Chat	1450±4	34316
79	Claude Opus 4.1 (Thinking 16K, 20250805)	1450±3	49535

code

核验来源

来源名次	模型（含配置）	原始分	样本数
1	Gemini 4 Carbon (High)	1552±15 Preliminary	1621
2	Gemini 4 Argon (High)	1550±10	4193
3	Claude Opus 4.7 (High)	1540±5	26193
4	Claude Fable 5 (High)	1538±6	16769
5	Claude Opus 5.5 (High)	1538±10	3876
6	Claude Opus 4.6 (High)	1537±5	31981
7	Muse Wonder (Max)	1535±12	2785
8	Claude Opus 4.6	1535±5	34720
9	Muse Spark 1.2 (xHigh)	1534±10	4340
10	GPT-6.1 Sol (Max)	1530±11	3156
11	Claude Opus 4.7	1530±5	26493
12	Claude Fable 5.1 (Max)	1529±9	5644
13	MiMo-V2.6-Pro	1528±14	1630
14	Muse Spark 1.3 (Max)	1525±7	8240
15	Muse Spark 1.1	1525±6	17897
16	Claude Opus 5 (High)	1524±5	27064
17	Kimi K3 (Max)	1523±6	13902
18	Claude Opus 4.8 (High)	1523±5	25108
19	Muse Spark	1522±8	5551
20	GPT-6 Astra (Max)	1522±9	5349
21	GPT-5.6 Sol (xHigh)	1522±6	16872
22	Claude Sonnet 5.5 (xHigh)	1519±11	3416
23	Gemini 3.8 Flash (High)	1518±6 Preliminary	17809
24	Claude Opus 5 (Max)	1518±6	12834
25	Claude Opus 4.8	1517±5	25737
26	GLM-5.3-Flash	1515±6	15570
27	DeepSeek-V4.1-Flash-20260910 (Max)	1514±9	5363
28	Claude Sonnet 4.6	1514±5	27747
29	GLM-5.3 (Max)	1513±7	9675
30	Gemini 3.1 Pro Preview	1513±4	51951
31	Qwen3.8-Max	1512±6	12546
32	Claude Opus 4.5 (High, 32K, 20251101)	1512±6	12636
33	Gemini 3.7 Flash (High)	1512±6 Preliminary	12292
34	GPT-5.5 (High)	1511±5	27468
35	Gemini 3.6 Flash (High)	1510±6	14504
36	GPT-5.2 Chat (20260210)	1510±6	14031
37	Gemini 3 Pro	1509±6	14390
38	Qwen3.7-Max-Preview	1509±15	1616
39	MiMo-V2.5-Pro	1509±5	30420
40	ERNIE 5.1	1508±6	15688
41	GPT-5.6 Terra (xHigh)	1507±6	18481
42	Claude Opus 4.5 (20251101)	1507±5	26861
43	GLM-5.2 (Max)	1507±5	20769
44	GPT-5.4 (High)	1507±5	25478
45	GPT-5.5	1506±5	28332
46	Claude Sonnet 5 (High)	1505±5	20391
47	GPT-6 Sol (Max)	1504±9	5692
48	Grok 4.20 Beta 1	1503±6	10827
49	Kimi K2.6	1503±6	16029
50	Grok 4.20 Beta (Reasoning, 0309)	1503±5	26051
51	GPT-5.5 Instant	1502±7	11165
52	Dola Seed 2.0 Pro	1502±5	30993
53	Grok 4.5	1502±6	15970
54	MiMo-V2.6-Flash	1502±9	5703
55	Grok 4.20 Multi-Agent Beta (0309)	1502±5	25579
56	GLM-5.1	1502±5	26126
57	Gemini 3.5 Flash (High)	1502±5	19402
58	Gemini 3 Flash	1500±7	10651
59	Qwen3.5-Max-Preview	1500±7	8991
60	Claude Sonnet 4.5 (High, 32K, 20250929)	1500±4	30407
61	GPT-5.4	1499±5	26996
62	Qwen3.6-Max-Preview	1497±13	2257
63	Claude Sonnet 4.5 (20250929)	1497±4	30303
64	Gemini 3.5 Flash (Medium)	1497±6	18867
65	Grok 4.1 (Thinking)	1496±5	23755
66	Grok 4.6 (High)	1493±6	12579
67	Hy3	1493±8	5452
68	GPT-5.6 Luna (xHigh)	1493±6	16040
69	DeepSeek-V4-Pro (High, 20260813)	1492±9	4093
70	Gemma 4 31B	1492±12	2217
71	Claude Opus 4.1 (Thinking 16K, 20250805)	1492±5	16722
72	DeepSeek-V4-Pro	1491±5	23647
73	GPT-6 Luna (Max)	1491±6	30405
74	Qwen3.7-Plus	1491±6	17157

image

核验来源

来源名次	模型（含配置）	原始分	样本数
1	GPT Image 2.5 Sunburst (max)	1222±10	13040
2	GPT Image 2.5 Flare (max)	1207±10	12380
3	GPT Image 2 (high)	1187±11	12857
4	Nano Banana 2.1	1176±12	6221
5	Grok Imagine Image 2.0	1166±12	7495
6	MAI-Image-2.6	1166±12	8353
7	Nano Banana 2 (Gemini 3.1 Flash Image)	1140±10	16088
8	Muse Image	1127±11	8491
9	GPT Image 1.5 (high)	1122±11	12845
10	MAI-Image-2.6-Flash	1119±12	5504
11	Nano Banana Pro (Gemini 3 Pro Image)	1118±10	15193
12	MAI-Image-2.5	1118±11	13804
13	FLUX 3 Image	1117±12	6137
14	MAI-Image-2.5-Pro	1114±10	12222
15	Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image)	1107±10	15091
16	Seedream 5.0 Pro	1101±11	11643
17	Qwen-Image-3.0-Pro	1096±12	6345
18	Qwen-Image-3.0	1070±12	6103
19	Qwen-Image-2.1-Pro	1057±12	4985
20	grok-imagine-image-quality	1049±9	16744
21	Qwen-Image-2.1	1043±12	5727
22	FLUX 3 Image Klein	1043±12	5188
23	Fibo 2.0	1039±12	5125
24	Krea 2 Large	1034±9	14520
25	Qwen Image 2.0 Pro (2026-04-22)	1030±11	6154
26	MAI-Image-2.5-Flash	1029±11	5875
27	Qwen-Image-2.1-Turbo	1029±12	4638
28	Wan 2.6 Image	1027±12	5382
29	Seedream 4.5	1026±12	6409
30	Recraft V4.1 Utility	1026±11	7471
31	FLUX.2 [flex]	1026±11	6091
32	Ideogram 4.0 (Quality)	1026±10	10708
33	Luma UNI 1 Max	1023±12	5549
34	grok-imagine-image	1023±11	6345
35	FLUX.2 [max]	1021±10	11113
36	Krea 2 Medium Turbo	1020±11	6075
37	Recraft V4.1 Utility Pro	1018±9	15426
38	Seedream 5.0 Lite	1017±12	5626
39	HiDream-O1-Image-1.5	1017±11	12182
40	Wan2.6 Text to Image	1015±12	5573
41	Ideogram 4.5 (High)	1015±15	2907
42	GPT Image 1 (high)	1013±12	5528
43	Krea 2 Medium	1012±9	14468
44	FLUX.2 [pro]	1003±10	10838
45	Ideogram 4.0	1002±10	10115
46	FLUX.2 [dev]	1000	18055
47	Ming-Image-0.1-Design	999±9	20252
48	P-Image-Ideogram (High)	995±14	3584
49	FLUX.2 [dev] Turbo	994±11	6287
50	Luma UNI 1	993±12	5571
51	Cosmos3-Super-Text2Image (agentic)	992±10	11566
52	P-Image-Ideogram (Very High)	991±15	2791
53	Qwen Image 2.0 (2026-03-03)	990±16	2645
54	FLUX.2 [dev] Flash	988±12	5353
55	HunyuanImage 3.0 (Tencent Cloud)	988±16	2544
56	HunyuanImage 3.0 Instruct	987±12	5696
57	Wan 2.7	986±12	5516
58	Cosmos3-Super-Text2Image	985±10	10900
59	Nano Banana (Gemini 2.5 Flash Image)	985±12	5326
60	Ideogram 4.0 Fast (Quality)	982±12	5960
61	Ideogram 4.0 Instant	982±12	5509
62	Wan 2.7 Pro	979±12	5513
63	HiDream-O1-Image	975±10	10674
64	P-Image-Ideogram (Medium)	972±14	3778
65	Cosmos3-Super-Text2Image-4Step	970±10	9044
66	HunyuanImage 3.5 (Preview)	970±12	5599
67	Ideogram 4.0 Fast	963±12	5454
68	P-Image-Ideogram (Low)	960±14	3844
69	Qwen Image Plus 2601	952±16	2596
70	Recraft V4.1 Flash	949±12	5785
71	GPT Image 1 Mini (medium)	936±17	2344
72	Z-Image Turbo	936±11	8187
73	HunyuanImage 3.0	932±16	2456
74	Dreamina 3.1	932±17	2383
75	FLUX.2 [klein] 9B	930±10	10747
76	FLUX.1 Kontext [max]	929±12	5294
77	FLUX.1 Kontext [pro]	918±18	2235
78	Kling Image 3.0 Omni	917±17	2690
79	ERNIE Image Turbo	915±10	10574
80	Ideogram 3.0	910±18	2345
81	P-Image-Ideogram (Very Low)	909±15	3326
82	ERNIE Image	907±10	10767
83	Qwen Image	903±18	2357
84	P-Image	899±18	2367
85	FLUX.1 Krea [dev]	897±39	529
86	Vidu Q2	895±17	2539
87	Z-Image Base	891±17	2524
88	GLM-Image	885±18	2476
89	Lucid Origin Ultra	878±17	2708
90	FLUX1.1 [pro]	870±19	2254
91	Lucid Origin Fast	860±18	2628
92	LongCat Image	857±19	2161
93	FLUX1.1 [pro] Ultra	855±19	2296
94	FIBO	849±19	2270
95	Luma Photon	844±19	2265
96	FLUX.2 [klein] 4B	837±12	6435
97	Phoenix 0.9 Ultra	834±20	2223
98	FLUX.1 [dev]	827±44	529
99	Phoenix 1.0 Fast	826±20	2169
100	SRPO	826±20	2166
101	FIBO Lite	813±16	3314
102	HiDream-O1-Image-Dev	811±20	2338
103	Luma Photon Flash	769±23	2065
104	FLUX.1 [schnell]	739±20	3060
105	OmniGen V2	717±26	2057
106	Bagel	594±35	2096

video

核验来源

来源名次	模型（含配置）	原始分	样本数
1	Wan 3.0	1173±14	6961
2	Utopai X	1168±14	6410
3	MiniMax H3 (768p)	1162±14	7102
4	FLUX 3 Max	1157±17	3390
5	HeyGen Video 1	1156±14	6142
6	Dreamina Seedance 2.5	1156±14	7093
7	MiniMax H3 Max	1148±17	3436
8	FLUX 3	1143±14	6117
9	Dreamina Seedance 2.0	1137±14	6096
10	Gemini Omni Flash 1.1	1130±13	7439
11	P-Video-2-Pro	1128±17	3061
12	Agnes-Video-2.5	1066±15	4876
13	Grok Imagine Video 1.5	1059±15	4988
14	HappyHorse-1.1	1050±14	6340
15	MAGI-2 Preview	1044±14	6590
16	Wan2.7-260612	1037±15	5618
17	HappyHorse-1.0	1033±15	4847
18	Kling 3.0 Omni 1080p (Pro)	1024±15	4598
19	Kling 3.0 1080p (Pro)	1000	5934
20	Grok Imagine Video 1.5 Lite	991±14	7199
21	PixVerse V6	984±16	4002
22	Veo 3.1	961±15	4229
23	Veo 3.1 Fast	957±15	5114
24	Veo 3.1 Lite	952±16	4090
25	LTX-2.5 Fast	947±16	4104
26	LTX-2.5 Pro	943±16	3946
27	Vidu Q3 Pro	937±15	4929
28	Vidu Q3 Turbo	915±16	4531
29	LTX-2.3 Pro	886±16	4587
30	LTX-2.3 Fast	839±17	4248
31	Boreal	832±17	3940

audio

核验来源

来源名次	模型（含配置）	原始分	样本数
1	Suno v6	1144±18	3037
2	Suno v6-mini	1132±18	3151
3	Mureka V9.5	1102±18	3107
4	Mureka V9	1063±18	3097
5	Lyria 3.5	1051±18	3164
6	Lyria 3 Pro	1048±18	3158
7	Eleven Music v2.5	1016±18	3220
8	Eleven Music v2	1002±18	3259
9	MiniMax Music 3.0	1000	3299
10	Eleven Music	999±18	3214
11	MiniMax Music 2.6	998±18	3206
12	MiniMax Music 2.5+	990±18	3337
13	Melodia v3	964±18	3293
1	Eleven v4 (TTS)	1333	未披露