Skip to content

Wan2.1-VACE-1.3B, Explained Simply

What Wan2.1-VACE-1.3B is, what it does well, and what it costs — released 2025-05-13 by Alibaba.

TLDR
  • Wan2.1-VACE-1.3B is Alibaba's new "all-in-one" AI model for creating and editing short videos.
  • It's a smaller, lightweight version designed to run on regular home computers, not just giant servers.
  • It can handle multiple jobs: text-to-video, image-to-video, and video editing, all in one model.

Want a heads-up when Wan updates?

We email you the moment a new Wan version drops — plus plain-English release notes like these. No spam.

Choose update emails Wan · Vibe Mastermind Updates
Choose which update emails you want

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.

Wan2.1-VACE-1.3B is a newly released video-generation and editing tool made by Alibaba. It's built for creators, developers, or students who want to make or tweak short videos using AI without needing super-expensive, high-end equipment.

What it does well

  • All-in-one design: Instead of needing one tool to make a video and another to edit it, VACE handles both creating and editing in a single model.
  • Runs on home GPUs: The 1.3B version is small enough to run on consumer-grade graphics cards (the kind found in gaming PCs). The larger Wan2.1 T2V-1.3B model needs only about 8.19 GB of VRAM (video memory) and can make a 5-second 480P video on an RTX 4090 in roughly 4 minutes.
  • Multiple tasks: It supports text-to-video, image-to-video, video editing, text-to-image, and video-to-audio.
  • Text in videos: Wan2.1 is the first video model that can generate both Chinese and English text inside the videos it creates.

How it compares

  • Vs. older Wan2.1 models: The original Wan2.1 models focused mostly on generating video. VACE is newer (announced May 14, 2025) and is built specifically as an "all-in-one" model for both video creation *and* editing.
  • Vs. closed-source rivals: The makers say Wan2.1's performance is comparable to some closed-source (paid/proprietary) models, while being open and available for people to download and use.

What it costs and what it can handle

  • Resolution: The VACE-1.3B model supports 480P resolution. (Note: The 1.3B model *can* technically generate 720P, but the makers warn it has limited training at that resolution, so results may not be as good.)
  • Hardware: Designed for consumer-grade GPUs. For context, an RTX 4090 can generate a 5-second 480P video in about 4 minutes.
  • Availability: The model is free to download from Hugging Face and ModelScope, and it works with tools like ComfyUI and Gradio.

Worth knowing

  • Lower quality at higher resolution: While it can attempt 720P, it wasn't trained on it much, so 480P is the safer choice for good results.
  • Smaller model, lower detail: As the 1.3B version, it's the lightweight option. There's also a larger VACE-14B model that handles both 480P and 720P better, but it requires more powerful hardware.
  • Still being integrated: Some software integrations (like Diffusers + Multi-GPU support) are still listed as "to do," so it may not work perfectly with every tool just yet.

This is our plain-language summary. Read the complete, official notes on the Wan official changelog ↗.

Vibe Mastermind in 00d 00h 00m 00s
Call LIVE — Join Now