Alibaba Launches WAN 3.0: 30-Second Video Generation with Multimodal References
aziz4ai · x · 2026-08-06
Alibaba's AI platform WAN has released the new video generation model WAN 3.0. The model excels in understanding physics, realism, and light/shadow details. It supports up to 30 seconds of video generation, setting a new industry standard. Furthermore, WAN 3.0 supports multimodal format references, allowing users to combine slides, Excel files, PDFs, alongside video, audio, and images as foundational references for creative generation.
Related event: Alibaba Releases Wan3.0 Video Model with Native 30-Second Generation(5 posts)→
More from Multimodal
- TwelveLabs Exec: Native Video Understanding is AI's Next Frontier — bigdata · 2026-08-06
- Hands-on with ByteDance's Seedance 2.5: Usable AI video in 1-2 tries, API coming soon — nikola_mr64990 · 2026-08-06
- Creating an AI Music Video "Fuzzy Wuzzy" Using MiniMax — Peemore · 2026-08-06
- MiniMax H3 Tested: Amazing Video Prompt Adherence — jefharris · 2026-08-06
- MiniMax H3 + AE Workflow: Run 40GB Models on 16GB VRAM for Commercials — circlenline · 2026-08-06
- Kimi K3 vs Grok 4.5: Which Model Better Completes Unfinished Sketches? — CodeByPoonam · 2026-08-06