Text-to-Video Arena

View overall rankings across text to video AI models.

May 12, 2026
420,893 votes
39 models
Rank Spread
1
11
Bytedance
Bytedance · Proprietary
1457±9
22,185
2
22
Alibaba-ATH · Proprietary
1435±9
6,266
3
38
Google · Proprietary
1372±11
13,978
4
38
OpenAI · Proprietary
1368±8
33,475
5
311
Google · Proprietary
1366±14
13,689
6
39
Google · Proprietary
1364±11
39,325
7
311
Google · Proprietary
1364±11
14,089
8
311
xAI · Proprietary
1357±7
108,058
9
512
Google · Proprietary
1349±11
25,154
10
612
Alibaba · Proprietary
1341±11
24,738
11
612
Google · Proprietary
1341±12
18,966
12
912
OpenAI · Proprietary
1339±7
44,913
13
1317
Alibaba · Proprietary
1260±13
13,064
14
1316
Bytedance
Bytedance · Proprietary
1258±7
60,453
15
1318
Google · Proprietary
1254±11
14,949
16
1318
Google · Proprietary
1250±12
15,230
17
1520
Pixverse · Proprietary
1238±9
20,975
18
1423
Runway · Proprietary
1235±12
20,839
19
1725
KlingAI · Proprietary
1221±17
2,104
20
1823
KlingAI · Proprietary
1219±7
60,034
21
1827
Pruna · Proprietary
1209±16
7,041
22
1829
Luma AI · Proprietary
1207±22
1,121
23
1730
KlingAI · Proprietary
1207±27
1,193
24
2028
MiniMax · Proprietary
1199±12
9,370
25
2028
MiniMax · Proprietary
1199±7
50,014
26
2129
Bytedance
Bytedance · Proprietary
1192±11
12,122
27
2231
MiniMax · Proprietary
1181±12
9,333
28
2131
Kandinsky
Kandinsky · MIT
1176±21
2,020
29
2431
Tencent
Tencent · tencent-hunyuan-community
1170±16
4,273
30
2631
Google · Proprietary
1164±16
6,509
31
2731
KlingAI · Proprietary
1164±9
14,049
32
3234
Lightricks
lightricks · ltx-2-community-license-agreement
1135±9
42,742
33
3235
Alibaba · Apache 2.0
1133±15
10,419
34
3235
Kandinsky
Kandinsky · MIT
1115±18
1,475
35
3335
Bytedance
Bytedance · Proprietary
1114±9
16,214
36
3637
OpenAI · Proprietary
1070±16
4,080
37
3637
Luma AI · Proprietary
1066±17
5,217
38
3839
Pika · Proprietary
1009±15
5,728
39
3839
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

Confidence Intervals on Model Strength (via Bootstrapping)

Battle Count for Each Combination of Models (without Ties)

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