Best AI Video Generators: Video Model Leaderboard
Ranked by LMArena scores from anonymous head-to-head human votes. Artificial Analysis Video Arena Elo is shown for comparison; the two scales use different baselines and cannot be compared directly.
- Top open-weight
MiniMax H31460MiniMax - #1 on AA (matched)
Wan 3.0AA Elo 1135Alibaba
On LMArena’s text-to-video leaderboard published Sep 22, 2026, #1 is Gemini Omni 1.1 Flash by Google (1516, 1,784 votes); the image-to-video leader is MiniMax H3 (1495).
Full leaderboard · Text-to-video
7 modelsSorted by LMArena score by default; ± is half the 95% confidence interval.
| # | Model | LMArena | Artificial Analysis | Released | License | |||
|---|---|---|---|---|---|---|---|---|
| Score↑ | Rank | Votes | Elo | Rank | ||||
| 1 | GMochi V1GenmoOpen | 1007±17 | 48 | 5,893 | – | – | – | Apache 2.0 |
| 2 | KKandinsky 5.0 T2V LiteKandinskyOpen | 1113±18 | 44 | 1,474 | – | – | – | MIT |
| 3 | Wan V2.2 A14BAlibabaOpen |
1132±15 | 42 | 10,458 | – | – | – | Apache 2.0 |
| 4 | LLTX 2 19BLightricksOpen | 1154±8 | 41 | 75,596 | – | – | – | ltx-2-community-license-agreement |
| 5 |  Hunyuan Video 1.5TencentOpen |
1169±16 | 38 | 4,292 | – | – | – | tencent-hunyuan-community |
| 6 | KKandinsky 5.0 T2V ProKandinskyOpen | 1175±21 | 37 | 2,019 | – | – | – | MIT |
| 7 | MiniMax H3MiniMaxOpen |
1460±9 | 8 | 11,417 | 1122±18 | 3 | Jul 2026 | minimax-h3-community-license-agreement |
Scores: LMArena leaderboard-dataset (lmarena-ai on Hugging Face, CC-BY-4.0), Text-to-video published Sep 22, 2026; AA columns from the Artificial Analysis Video Arena. Please credit the sources when citing.
Sources & methodology
- LMLMArena leaderboard dataset (Hugging Face, CC-BY-4.0)
published Sep 22, 2026 · 48 models
- AAArtificial Analysis Video Arena
fetched Oct 26, 2026 · 32 models
- Scores come from LMArena’s text_to_video and image_to_video leaderboards (overall category, latest release): users vote between two anonymous videos and LMArena fits a Bradley–Terry model.
- AA values are shown only when the model name (ignoring bracketed variant notes) maps to exactly one row; hover to see the AA entry name. Anything unmatched stays blank (–); nothing is guessed.
- Price, clip length, resolution and speed are not provided by LMArena or the AA API, so they are not shown; 720p/1080p in a name is the variant LMArena tested.
- Synced daily at 07:10 UTC; public JSON: /api/public/video-leaderboard.php
This week
Collecting data: 1/7 days recorded; weekly changes appear after 7 days
Synced automatically once a day; ranks are compared only between real snapshots, never backfilled.FAQ
What is the best AI video generator right now?
On LMArena’s text-to-video leaderboard published Sep 22, 2026, #1 is Gemini Omni 1.1 Flash by Google (1516, 1,784 votes); the image-to-video leader is MiniMax H3 (1495).
Which vendors rank highest for video generation?
On the text-to-video board, each vendor's best model ranks: Google Gemini Omni 1.1 Flash (1516, #1), Black Forest Labs FLUX 3 Video (1493, #3), xAI Grok Imagine Video 1.5 Agent (1492, #4), ByteDance Dreamina Seedance 2.0 720p (1479, #5), Alibaba Wan 3.0 (1476, #6).
Which open-weight video models can I run locally?
7 text-to-video models have non-proprietary licenses; the highest is MiniMax H3 (1460, license minimax-h3-community-license-agreement). Community licenses may restrict commercial use; check the license text.
Why do LMArena and AA scores differ so much?
The two platforms use different voters, prompts and baselines, so absolute numbers are not comparable; relative order within one platform is what matters.