Skip to content

Vchitect 2.0

Generate 20-second videos and evaluate long-video models

Visit official site

What Vchitect 2.0 is

The model is described as an advanced video generation model developed by Shanghai AI Laboratory, supporting 20-second video generation, flexible aspect ratios, spatial-temporal enhancement and long-video evaluation. It can turn static images into 5 to 10 second videos, and the VBench platform provides a comprehensive, continuously updated leaderboard covering long-video models such as Gen-3, Kling and OpenSora, with an open-source one-click evaluation deployment.

What you can do with it

  • Generate 20-second videos with flexible aspect ratios
  • Apply spatial-temporal enhancement during generation
  • Convert static images into 5 to 10 second videos
  • Evaluate long-video generation models through VBench

Application scenarios

  • Turn static images into dynamic videos for social media promotion
  • Enrich online courses and tutorials with generated video content
  • Compare video generation models with the VBench platform

Getting started

  • Visit the official Vchitect 2.0 website
  • Choose Text-to-Video or Image-to-Video as needed
  • Upload the corresponding text or image file
  • Set the aspect ratio and other generation parameters

Pros & cons

✓ What we like

  • Spatial-temporal enhancement in generation
  • Evaluation covers Gen-3, Kling and OpenSora among long-video models

! What to watch out for

  • Image-to-video output is described as 5 to 10 seconds long
  • Model evaluation runs through the VBench platform

FAQ

How long can generated videos be?

The material states 20-second video generation, and static images can be turned into 5 to 10 second videos.

What is VBench used for?

It supports evaluation of long-video generation models with a comprehensive and continuously updated leaderboard.

Last reviewed: 2026-11-15

More AI video generation tools

View all →

How we review