
What happened
Wan 3.0 took #1 on Artificial Analysis' text-to-video leaderboard with 1,174 Elo, ahead of Runway Gen-5 at 1,169 and Google Veo 3.1 at 1,165.
Why it matters
The top ranking signals that Alibaba's model is now competitive at the front of the field for text-to-video generation, giving teams a credible new option to evaluate.
Summaries like this, every morning.
Wan 3.0's rise to the top of the Artificial Analysis text-to-video leaderboard comes with a narrow margin: its 1,174 Elo sits just five points above Runway Gen-5 and nine above Google Veo 3.1. That tight cluster suggests the leading models are separated by small differences in blind human evaluations rather than a single dominant system.
The broader appeal of Wan 3.0 lies in how it handles long clips. The model accepts up to ten reference images, five videos, five audio clips and a document of up to 100MB, and it lets users define camera moves in five-second time codes across shots. This lets teams feed a planning document, a start frame and a motion reference directly into the generation process, a more reliable route to a coherent 30-second brand video.
Cost control is part of the pitch. At 720p, generation costs 5 credits per second, so a 5-second test uses 25 credits and a 30-second draft uses 150 credits. The 1080p rate of 15 credits per second brings a 30-second master to 450 credits. Testing at 720p and 5 seconds first avoids wasting credits on failed long takes.
Pick your industry and the AI tools you use, and get news related to your work every day.
Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.