AlibabaAlibaba·🎬 Video Generation·New

Wan 3.0 Pro Reference

anonymized
Try on Venice.ai ↗
Quick reference
Wan 3.0 Pro Reference — TLDR
  • 🏢 Alibaba's Wan 3.0 video family, Pro tier, reference-driven generation.
  • 🆕 Reference-to-video mode: images, video, audio or documents condition the output.
  • 📏 Up to 30 seconds in a single generation, 30fps, up to 1080p.
  • 🔧 All-in-One design also covers text-to-video and first/last-frame control.
  • 💬 Native audio in the same pass — dialogue, background music, sound effects.
  • 🎯 Built for subject consistency across frames from reference assets.
  • 📚 Accepts up to 20 multimodal reference materials per request.
  • 🔒 Access is via Alibaba Cloud Model Studio's hosted API.
💰 Pricing
$0.780 – $12.68
per generation
📅 On Venice since
Sep 8, 2026
1 day ago
Provider

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,…

Read full profile →
74 models on Venice
35 video · 22 text · 7 image · 6 inpaint · 2 embedding · 2 tts
Since Jan 11, 2025

About this model

Wan 3.0 Pro Reference is the reference-conditioned entry point to the Pro tier of Alibaba's third-generation Wan video family. Instead of starting from a prompt alone, you supply reference material — images, clips, audio, or even documents and webpages — and the prompt specifies how those reference assets are combined into a single shot. It sits alongside Wan 3.0 Pro text-to-video and Wan 3.0 Pro image-to-video, which share the same generation stack.

The generational jump from Wan 2.7 Reference is largely about scale and unification. Alibaba describes the Wan 3.0 series as an "All-in-One" model with comprehensive upgrades in audio-video generation, multimodal reference, and video editing, supporting up to 30 seconds per generation at 30fps, native dialogue, background music and sound effects, and up to 20 reference materials in a single request. One model name covers text-to-video, first-frame/last-frame image-to-video, and reference-based generation without switching endpoints.

Within Wan 3.0 there are tiers: the base Wan 3.0 Reference, this Pro variant, and the higher-capability Wan 3.0 Prime Reference and Wan 3.0 Prime Pro Reference lines. This model carries a catalog release date of September 8, 2026.

Access is through Alibaba Cloud Model Studio's hosted video generation API, which handles asynchronous task submission and result retrieval. Typical applications include multi-shot narrative clips, character-consistent series content, and turning existing assets into finished footage.

This About section is AI-generated from public sources (Claude Opus 5), with no human editing. It may contain inaccuracies — verify critical details against the sources listed above.

Data sources: Venice API · HuggingFace · Wikipedia — enrichment updated 14h ago