Breaking

Alibaba's Wan 3.0 Tops Artificial Analysis's Updated Text-to-Video Benchmark

October 1, 2026 at 07:29 EDT

Media Generation
Foundation Models

Alibaba's Wan 3.0 has taken the No. 1 overall slot on Artificial Analysis's updated AA-Video-T2V v2.0 leaderboard, scoring 1,156 Elo for text-to-video generation with audio. The lead is meaningful but not runaway: the top four entries sit within 17 Elo points, and Wan 3.0's 95% interval overlaps the next-ranked Utopai X.

Aug. 2026 release · Alibaba

Wan 3.0 takes the video lead

Alibaba's model tops the updated text-to-video ranking and also leads video editing, while the image-to-video crown remains with MiniMax H3 Max.

Text-to-video race: Elo above the 1,000 anchor

Wan 3.0

1,156 Elo

Utopai X

1,148 Elo

Seedance 2.5

1,139 Elo

MiniMax H3

1,139 Elo

The top four are close: Wan 3.0 leads by 8 Elo over Utopai X and by 17 Elo over the next two entries.

Overall text-to-video rank

6,738

T2V samples

30s

Maximum generation length

$12

Listed API price per minute

One model, several production paths

Input

Text, images, audio, video, files or links

→

Control

References, first frame or first-last frame

→

Output

MP4 video with audio on by default

Why creators care

Longer single-pass clips, document-to-video input and multi-reference control point toward ads, demos, training clips and short narrative beats.

What to watch

The text-to-video lead is narrow, image-to-video is led by another model, and API cost and regional constraints still matter.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.