Text-to-Video Arena

View overall rankings across text to video AI models.

May 29, 2026
452,799 votes
40 models
Rank Spread
1
11
Bytedance
Bytedance · Proprietary
1463±9
29,002
2
22
Alibaba-ATH · Proprietary
1440±8
11,701
3
39
Alibaba · Proprietary
1385±21
1,153
4
310
Google · Proprietary
1370±11
16,707
5
310
OpenAI · Proprietary
1367±8
35,490
6
311
Google · Proprietary
1366±14
13,688
7
310
Google · Proprietary
1364±11
39,330
8
311
Google · Proprietary
1362±11
16,837
9
311
xAI · Proprietary
1359±7
121,372
10
413
Google · Proprietary
1349±11
25,154
11
713
Google · Proprietary
1341±12
18,967
12
1013
Alibaba · Proprietary
1339±12
27,529
13
1013
OpenAI · Proprietary
1338±7
47,593
14
1419
Alibaba · Proprietary
1259±13
15,751
15
1417
Bytedance
Bytedance · Proprietary
1259±7
63,894
16
1419
Google · Proprietary
1254±11
14,949
17
1419
Google · Proprietary
1249±12
15,230
18
1521
Pixverse · Proprietary
1239±9
23,536
19
1524
Runway · Proprietary
1237±13
23,472
20
1826
KlingAI · Proprietary
1221±17
2,104
21
1924
KlingAI · Proprietary
1219±7
63,012
22
1928
Pruna · Proprietary
1210±16
7,040
23
1831
KlingAI · Proprietary
1207±27
1,193
24
1930
Luma AI · Proprietary
1207±22
1,121
25
2129
MiniMax · Proprietary
1199±7
53,264
26
2130
MiniMax · Proprietary
1199±12
9,370
27
2230
Bytedance
Bytedance · Proprietary
1192±11
12,121
28
2332
MiniMax · Proprietary
1181±12
9,333
29
2232
Kandinsky
Kandinsky · MIT
1176±21
2,020
30
2432
Tencent
Tencent · tencent-hunyuan-community
1170±16
4,273
31
2732
Google · Proprietary
1164±16
6,509
32
2832
KlingAI · Proprietary
1164±9
14,049
33
3335
Lightricks
lightricks · ltx-2-community-license-agreement
1137±9
45,852
34
3336
Alibaba · Apache 2.0
1133±15
10,418
35
3336
Kandinsky
Kandinsky · MIT
1115±18
1,475
36
3436
Bytedance
Bytedance · Proprietary
1114±9
16,214
37
3738
OpenAI · Proprietary
1070±16
4,080
38
3738
Luma AI · Proprietary
1066±17
5,217
39
3940
Pika · Proprietary
1009±15
5,728
40
3940
Genmo AI · Apache 2.0
1007±16
5,862

Remove Style Control Leaderboard Plots

Fraction of Model A Wins for All Non-tied A vs. B Battles

Battle Count for Each Combination of Models (without Ties)

Average Win Rate Against All Other Models (Uniform Sampling and No Ties)

Confidence Intervals on Model Strength (via Bootstrapping)