Alibaba's latest AI video model in the Tongyi Wanxiang (Wan) line. It generates clips up to about 30 seconds — roughly double the prior Wan 2.x generation — from text, images, or documents like PDFs and slide decks. Fully launched August 24, 2026.
Last verified · 2026-08-25 · by Moe Ameen
Wan3.0 is Alibaba's latest AI video generation model, part of its Tongyi Wanxiang (Wan) family. Alibaba rolled it out to a full public launch on August 24, 2026, after opening a public beta earlier that month. The headline capability is length: Wan3.0 generates clips up to about 30 seconds — roughly double the ~15-second ceiling of the previous Wan 2.x generation — while holding character detail, spatial layout, and on-screen motion graphics steady across the full clip rather than drifting after the opening seconds. Reported output runs to 1080p.
What separates it from a plain text-to-video model is multimodal input. Alongside text and images, Wan3.0 accepts video, audio, and — most notably — documents: web pages, PDFs, slide decks, and spreadsheets. You can hand it a marketing one-pager or a set of slides and get a narrated, animated video back, which is why Alibaba has aimed the model at short-drama and film production, advertising and marketing, tourism promotion, and music-video creation. It also generates multilingual voice and realistic facial expressions in the clip.
Access at launch is hosted, not a download. Wan3.0 is reached through Alibaba Cloud's Model Studio and Qwen platforms, with beta access granted by application. Alibaba Cloud lists usage-based API pricing for Wan3.0 on Model Studio — roughly $0.05, $0.10, and $0.20 per second of generated video for 480p, 720p, and 1080p respectively, so a full 30-second 1080p clip runs about $6. Earlier Wan releases shipped open weights, but treat Wan3.0's own weight availability as unconfirmed at the time of writing. The release landed a day after Alibaba raised roughly $10 billion in a share placement earmarked for its AI buildout, so expect the model and its distribution to move fast.
Wan3.0's most interesting trick is turning a document — a slide deck, a PDF, a one-pager — into a single polished 30-second video. That is genuinely new, but notice what you have at the end: one clip. A 30-second video is a great asset and a terrible content calendar. This is where [Kompozy](/) does the other half of the job. Feed that Wan3.0 clip in as a source and Kompozy fans it into roughly 25–35 assets across 18 formats: [Clipped Shorts](/glossary/content-repurposing) that cut the 30 seconds into three or four separate hooks, a captioned [Persona Short](/glossary/persona-shorts) that reframes the message with a face-locked avatar, a brand-exact [Carousel](/glossary/hyperframes), quote graphics, a blog article, and an email newsletter — the same source deck now working as a week of posts instead of one upload.
The two things Wan3.0 does not touch are exactly what Kompozy is for. Identity: a [Persona Brief](/glossary/persona-brief) and an AI Influencer persona pool with Gemini face-lock keep the same voice and recognizable presenter across every video and image, so a folder of Wan3.0 clips reads as one brand rather than a pile of one-offs. Distribution: Wan3.0 hands you a file gated behind an Alibaba Cloud application, while Kompozy's [Autopilot](/glossary/autopilot) captions, resizes per platform, and schedules the whole set across the eight social platforms plus blog and email, each asset clearing a per-post review gate first. Wan3.0 makes the 30-second video; Kompozy makes it a month of published content.
Wan3.0 is Alibaba's latest AI video generation model, part of its Tongyi Wanxiang (Wan) line. It generates clips up to about 30 seconds — roughly double the previous Wan 2.x generation — from text, images, or documents such as PDFs and slide decks, and holds character, layout, and motion graphics steady across the full clip. Alibaba fully launched it on August 24, 2026.
Alibaba rolled Wan3.0 out to a full public launch on August 24, 2026, following a public beta earlier that month. Access at launch is hosted through Alibaba Cloud's Model Studio and Qwen platforms, with beta access granted by application rather than an open consumer app.
Two things: length and inputs. It generates clips up to about 30 seconds — double the prior Wan generation — and it accepts documents (web pages, PDFs, slide decks, spreadsheets) alongside text, images, video, and audio, so a deck or one-pager can become a narrated video. It also renders multilingual voice and facial expressions in the clip.
Wan3.0 is not free: it is a hosted model reached through Alibaba Cloud Model Studio and Qwen, billed by usage at about $0.05, $0.10, and $0.20 per second of video for 480p, 720p, and 1080p (roughly $6 for a 30-second 1080p clip). Earlier Wan releases shipped open weights, but Wan3.0's open-weight availability was unconfirmed — confirm current pricing and access on Alibaba Cloud before relying on it.
Bring the clip into Kompozy. From that single source Kompozy cuts the 30 seconds into several captioned shorts and fans the idea into a carousel, quote graphics, a persona/avatar cut, a blog, and a newsletter — holds a consistent face and voice across them, and schedules and publishes the set across the eight social platforms plus blog and email.