Back to products
Odyssey-2 Max

Odyssey-2 Max

Physical accuracy takes a leap in world models Artificial Intelligence • 3D Modeling • Video 2 126 Framer Launch websites with enterprise needs at startup speeds. Design Tools • Website Builder • Artificial Intelligence

Overview

What it is

Odyssey is a new AI lab launching a research preview of real-time interactive video. Powered by a world model, it generates explorable video experiences you can influence.

Intent

I need it when

Create interactive, multi-minute video simulations that users can control and explore in real-time

Odyssey-2 Max streams interactive simulations instantly (50ms generation) with minutes of content and open-ended interaction, unlike traditional video models that generate fixed 10-second clips over minutes

Develop edge-case scenario generation and realistic training simulations for high-stakes domains

Odyssey-2 Max generates controllable, realistic edge-case scenarios and simulated training environments for warfighters and other specialized training needs with physical accuracy

Build intelligent interactive applications for gaming, retail training, fitness, and customer support

Odyssey-2 Max powers emergent media, machine training, companionship, and intelligent assistance use cases including interactive games, staff training, personalized tutoring, and empathetic customer support

Generate realistic, interactive world simulations from text or image prompts for robotics training and task rehearsal

Odyssey-2 Max enables robots to simulate complex manipulation, reaching, and navigation tasks learned from video before real-world deployment, reducing brittleness and enabling adaptation to real-world variability

Drop

Not a fit when

  • User needs fixed-duration video generation; Odyssey-2 Max generates multi-minute interactive simulations requiring real-time interaction
  • User requires offline-only operation; the product appears to be API-based and cloud-dependent
  • User needs deterministic, non-interactive outputs; Odyssey-2 Max is designed for open-ended, user-driven interaction
  • User operates in highly constrained environments with no video data; the model learns from decades of video and requires visual input
  • User needs immediate production-ready solutions; Odyssey-2 Max is described as early-stage technology
Commercials

Pricing

Pricing not specified