Alibaba leads artificial intelligence video rankings with Wan 3.0 model

Executive Briefing
“Alibaba's Wan 3.0 model secured the top position on the Artificial Analysis text-to-video...”
Key Takeaways
- 1.Alibaba's Wan 3.0 achieved 1175 Elo on the text-to-video leaderboard at a price of $0.20 per second.
- 2.ByteDance Seedance 2.5 led the image-to-video category with 1161 Elo but carried the highest cost on the board.
- 3.MiniMax H3 ranked third as an open-weights model with 1161 Elo while serving as a fine-tuning base for several other variants.
Alibaba's Wan 3.0 model secured the top position on the Artificial Analysis text-to-video leaderboard with 1175 Elo points. Meanwhile, open-weights models like MiniMax H3 formed the foundation for multiple high-performance variants.
Key Developments and Technical Context
Alibaba's Wan 3.0 achieved 1175 Elo on the text-to-video leaderboard at a price of $0.20 per second.
ByteDance Seedance 2.5 led the image-to-video category with 1161 Elo but carried the highest cost on the board.
MiniMax H3 ranked third as an open-weights model with 1161 Elo while serving as a fine-tuning base for several other variants.
Strategic Analysis and Industry Implications
The emergence of multiple fine-tuned models from a shared open-weights foundation highlights a trend toward cost optimization and faster deployment in the artificial intelligence video ecosystem.
Outlook and Key Factors to Watch
The actual performance of models absent from blind-evaluation leaderboards such as OpenAI Sora 2 or Kling 4.0 still requires further verification.







