Welcome to the World Model Evaluation Study

Evaluate AI-generated world models

You will be evaluating AI-generated world models. Each evaluation involves comparing two videos side-by-side and determining which one performs better across several quality dimensions.

What You'll Be Doing:

  1. Watch two short videos (approximately 4 seconds each) side by side
  2. Compare them across four dimensions:
    • Overall Quality - General realism and believability
    • Controllability - How well the character follows commands
    • Visual Quality - Clarity and aesthetic appeal
    • Temporal Consistency - Smoothness and stability over time
  3. Select which video performs better for each dimension

Important Notes:

  • Each evaluation takes approximately 2-3 minutes
  • Please watch both videos completely before making your judgments
  • There are no right or wrong answers - we want your honest opinion
  • Use a modern browser (Chrome, Firefox, Safari) for best experience

Available Evaluations