Skip to content

Pandora

A world model controlled by free-text actions

Visit official site

What Pandora is

The model is described as a step toward a general world model that simulates world states by generating video.

What you can do with it

  • Control video with text actions
  • Generate across many domains
  • Simulate alternative futures

Application scenarios

  • Build dynamic game scenes
  • Preview script variations
  • Make animation layouts

Pros & cons

✓ What we like

  • Action control can transfer to unseen domains
  • FLAVR frame interpolation is listed

! What to watch out for

  • Control depends on text action instructions
  • Video length follows the autoregressive model

FAQ

How is video controlled?

The material states free-text action input during generation.

What scenes are named?

The material lists indoor, outdoor, natural, urban, human, robot, 2D and 3D scenes.

Last reviewed: 2026-11-15

More AI video generation tools

View all →

How we review