---
title: "Best AI for video generation"
description: "Video models ranked on the LMArena text to video, image to video and video edit arenas, with coverage shown for every row."
url: "https://aldena.ai/best-ai-for-video-generation"
---

# Best AI for video generation

Human preference over generated clips, from a prompt, from a still, and from an edit.

Aldena runs these models inside your team rooms. [See what each one costs](https://aldena.ai/models).

| rank | model | vendor | score | coverage (benchmarks) | LMArena text to video (elo) | LMArena image to video (elo) | LMArena video edit (elo) |
| --- | --- | --- | --- | --- | --- | --- | --- |
| 1 | Minimax H3 | MiniMax | 77.7 | 3 of 3 | 1453 | 1489 | 1390 |
| 2 | Dreamina Seedance 2.0 720p | ByteDance | 75.3 | 3 of 3 | 1478 | 1479 | 1358 |
| 3 | Gemini Omni Flash | Google | 73.0 | 3 of 3 | 1512 | 1462 | 1358 |
| 4 | FLUX.3 Video | Black Forest Labs | 71.5 | 2 of 3 | 1496 | 1453 | — |
| 5 | HappyHorse 1.0 | Alibaba | 66.3 | 3 of 3 | 1428 | 1441 | 1307 |
| 6 | Veo 3.1 Audio | Google | 65.6 | 2 of 3 | 1364 | 1397 | — |
| 7 | Veo 3.1 Audio 1080p | Google | 64.4 | 2 of 3 | 1363 | 1390 | — |
| 8 | Muse Video (partial coverage) | Meta | 63.6 | 1 of 3 | 1457 | — | — |
| 9 | Grok Imagine Video 1.5 720p (partial coverage) | xAI | 63.6 | 1 of 3 | — | 1460 | — |
| 10 | Veo 3.1 Fast Audio | Google | 63.2 | 2 of 3 | 1362 | 1384 | — |
| 11 | Grok Imagine Video 720p | xAI | 63.2 | 2 of 3 | 1347 | 1416 | — |
| 12 | Veo 3.1 Fast Audio 1080p | Google | 61.4 | 2 of 3 | 1358 | 1371 | — |
| 13 | Wan2.7 I2V (partial coverage) | Alibaba | 61.3 | 1 of 3 | — | 1426 | — |
| 14 | Sora 2 Pro (partial coverage) | OpenAI | 61.3 | 1 of 3 | 1366 | — | — |
| 15 | Veo 3 Fast Audio | Google | 58.5 | 2 of 3 | 1347 | 1324 | — |
| 16 | Grok Imagine Video 480p (partial coverage) | xAI | 57.3 | 1 of 3 | — | 1384 | — |
| 17 | Veo 3 Audio | Google | 56.7 | 2 of 3 | 1339 | 1330 | — |
| 18 | Wan2.7 T2V (partial coverage) | Alibaba | 55.7 | 1 of 3 | 1343 | — | — |
| 19 | Vidu Q3 Pro (partial coverage) | Shengshu | 55.7 | 1 of 3 | — | 1362 | — |
| 20 | Kling v3 Pro (partial coverage) | Kling AI | 54.9 | 1 of 3 | — | 1356 | — |
| 21 | Sora 2 (partial coverage) | OpenAI | 54.9 | 1 of 3 | 1340 | — | — |
| 22 | Wan2.6 T2V (partial coverage) | Alibaba | 53.3 | 1 of 3 | 1333 | — | — |
| 23 | Seedance v1.5 Pro | ByteDance | 53.1 | 2 of 3 | 1256 | 1307 | — |
| 24 | Wan2.5 I2V Preview (partial coverage) | Alibaba | 52.5 | 1 of 3 | — | 1321 | — |
| 25 | Wan2.6 I2V (partial coverage) | Alibaba | 51.7 | 1 of 3 | — | 1310 | — |
| 26 | Wan2.5 T2V Preview (partial coverage) | Alibaba | 50.9 | 1 of 3 | 1249 | — | — |
| 27 | Pixverse v5.6 | PixVerse | 50.1 | 2 of 3 | 1240 | 1299 | — |
| 28 | Veo 3 | Google | 48.9 | 2 of 3 | 1253 | 1257 | — |
| 29 | Gen-4.5 (partial coverage) | Runway | 48.6 | 1 of 3 | 1223 | — | — |
| 30 | Kling 2.5 Turbo 1080p | Kling AI | 47.7 | 2 of 3 | 1219 | 1275 | — |
| 31 | Kling 2.6 Pro | Kling AI | 47.7 | 2 of 3 | 1217 | 1293 | — |
| 32 | Veo 3 Fast | Google | 47.1 | 2 of 3 | 1248 | 1256 | — |
| 33 | Grok Imagine Video (partial coverage) | SpaceXAI | 47.0 | 1 of 3 | — | — | 1263 |
| 34 | Vidu Q2 Turbo (partial coverage) | Shengshu | 44.6 | 1 of 3 | — | 1243 | — |
| 35 | Hailuo 2.3 | MiniMax | 44.2 | 2 of 3 | 1203 | 1261 | — |
| 36 | Seedance v1 Pro | ByteDance | 43.6 | 2 of 3 | 1190 | 1272 | — |
| 37 | Kling O3 Pro (partial coverage) | Kling AI | 42.2 | 1 of 3 | — | — | 1255 |
| 38 | Kling v2.1 Standard (partial coverage) | Kling AI | 42.2 | 1 of 3 | — | 1228 | — |
| 39 | Kandinsky 5.0 T2V Pro (partial coverage) | Sber AI | 41.4 | 1 of 3 | 1172 | — | — |
| 40 | Ray 3 | Luma AI | 40.6 | 2 of 3 | 1205 | 1225 | — |
| 41 | Hailuo 02 Pro | MiniMax | 40.6 | 2 of 3 | 1198 | 1228 | — |
| 42 | Vidu Q2 Pro (partial coverage) | Shengshu | 39.8 | 1 of 3 | — | 1222 | — |
| 43 | Hailuo 02 Fast (partial coverage) | MiniMax | 38.3 | 1 of 3 | — | 1193 | — |
| 44 | Kling O1 Pro | Kling AI | 38.2 | 2 of 3 | 1205 | — | 1200 |
| 45 | Kling v2.1 Master | Kling AI | 37.6 | 2 of 3 | 1162 | 1234 | — |
| 46 | Hailuo 02 Standard | MiniMax | 37.6 | 2 of 3 | 1180 | 1223 | — |
| 47 | Kandinsky 5.0 T2V Lite (partial coverage) | Sber AI | 36.7 | 1 of 3 | 1113 | — | — |
| 48 | Hunyuan Video 1.5 | Tencent | 35.2 | 2 of 3 | 1169 | 1196 | — |
| 49 | Sora (partial coverage) | OpenAI | 35.1 | 1 of 3 | 1069 | — | — |
| 50 | Runway Gen4 Turbo (partial coverage) | Runway | 33.5 | 1 of 3 | — | 1051 | — |
| 51 | Mochi v1 (partial coverage) | Genmo | 32.7 | 1 of 3 | 1005 | — | — |
| 52 | Runway Gen4 Aleph (partial coverage) | Runway | 32.7 | 1 of 3 | — | — | 1183 |
| 53 | Veo 2 | Google | 32.3 | 2 of 3 | 1164 | 1165 | — |
| 54 | Wan v2.2 A14b | Alibaba | 31.1 | 2 of 3 | 1131 | 1169 | — |
| 55 | Seedance v1 Lite | ByteDance | 30.5 | 2 of 3 | 1112 | 1184 | — |
| 56 | Ltx 2 19B | Lightricks | 30.5 | 2 of 3 | 1151 | 1153 | — |
| 57 | Ray2 | Luma AI | 26.9 | 2 of 3 | 1064 | 1107 | — |
| 58 | Pika v2.2 | Pika | 25.1 | 2 of 3 | 1008 | 996 | — |

## How this ranks

Every benchmark value becomes a percentile among the models that have it, so accuracy scores, Elo ratings and word error rates compare without hand-tuned scaling. Metrics where lower is better are inverted first. Raw values are never summed or averaged across benchmarks. A model's mean percentile is then shrunk toward the mean of the models that were broadly benchmarked, so a model tested twice cannot outrank a broadly tested one on two lucky results. Turning a data source off runs that same ranking code again in your browser over the sources you left on.

A model scored on fewer than 2 of the 3 ranked benchmarks in this category still ranks here, on the benchmarks it does have, and its row carries a partial coverage mark. On an equal score it sits under the model that earned the same number across more of the board.

## Data sources

- [LMArena](https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset): CC BY 4.0. Arena ratings by LMArena, from the public leaderboard dataset.