Wan 3.0
alibaba/wan-3.0/text-to-videowan-3-0-text-to-video- 🆕 Alibaba's newest Wan video model on this catalog, August 2026
- 📏 Alibaba states up to 30 seconds of video per generation
- 🌐 Positioned by Alibaba as generating video from any input
- 🔧 This entry is the text-to-video path of the Wan 3.0 release
- 🎯 Photorealistic frames with strong subject coherence across shots
- 🏢 Delivered through Alibaba Cloud Model Studio video generation APIs
- 👁️ Sibling variants cover image-to-video and reference-to-video workflows
- 📚 No public parameter count, license, or weight release listed
Alibaba Group is a Chinese multinational technology company founded in 1999 and headquartered in Hangzhou, Zhejiang. Originally built around e-commerce and cloud computing, Alibaba has become one of the most prolific contributors to open-weight AI research, developing the Qwen…
Explore 51 more models by Alibaba Group →Wan 3.0 is the text-to-video entry point into the newest generation of Alibaba's Wan visual model line, which is developed alongside — but separate from — the Qwen language models such as Qwen 3.8 27B. On this catalog it is described as a photorealistic generator with strong subject coherence across frames, and it is dated August 2026.
The headline change versus earlier releases in this family, Wan 2.7 and Wan 2.6, is clip length and input breadth: Alibaba presents Wan 3.0 as producing up to 30 seconds of video in a single generation, and as accepting a wide range of input types rather than prompts alone. Longer single-pass output matters because it reduces the need to stitch several short clips together to build a continuous shot.
Access is through Alibaba Cloud Model Studio, whose video generation service exposes text-driven and image-driven synthesis endpoints. Alibaba does not publish a parameter count or weights for this release, so it is used as a hosted service rather than run locally.
Companion Wan 3.0 variants on this catalog handle the other input paths: Wan 3.0 image-to-video animates a supplied still, while Wan 3.0 Reference conditions generation on reference material. For downloadable Wan models, the catalog also lists Wan 2.2 Enhanced.
This About section is AI-generated from public sources via VeniceStats + Venice inference, with no human editing. It may contain inaccuracies.
Daily usage · Wan 3.0 · Organic sellers
Top sellers · Wan 3.0 · Organic
- Open Forge73.9%
- Venice.ai Proxy24.1%
- antseed-tidal-falcon-88cd1.4%
- ▲ Apex Ant0.4%
- attacompute.ai0.1%
| Seller | Health | Rep↓ | Sales | Video | Routing | $ / s | $ / video |
|---|---|---|---|---|---|---|---|
| Open Forge 0x1d90…b0aa | healthy | 91.11 | $213.20 | 36.1 min | #1 | $0.10 | — |
| ▲ Apex Ant 0x73b4…e736 | — | 81.71 | $1.23 | 27 s | #3 | $0.0585 | — |
| attacompute.ai 0xd5e7…7be0 | — | 51.22 | $0.42 | 5 s | gated | $0.1625 | — |
| antseed-tidal-falcon-88cd 0x80d1…88cd | — | 12.78 | $4.03 | 25 s | gated | $0.057 | — |
| OpenCheap 0x0ee1…6c41 | — | 12.43 | — | — | gated | $0.095 | — |
| Venice.ai Proxy 0x1f22…18c9 | failingupstream error · 3.7 h | 86.84 | $69.64 | 18.1 min | #2 | $0.0715 | — |
"Best price" and the seller table are live Antseed catalog data (advertised $/1M tokens — or $ per generated image for unit-billed image models — not settled amounts). Reputation = buyer trust score (0-100, the Antseed SDK's own formula). "Routing" = the SDK's default buyer routing (what the VPR desktop app ships with): a trust ≥ 60 gate on the effective reputation, then cheapest-first among routable sellers; live failover state (per-peer cooldowns) is buyer-side runtime and not included. Model knowledge (TLDR, provider, About) via the VeniceStats enrichment layer. Advertised catalog, not the model used in any specific purchase. "Usage on Antseed" counts only settlements whose buyers share the per-model split on-chain (metadata v2/v3, opt-in), so every usage figure is a lower bound.