Odyssey-2 Pro: The GPT-2 moment for interactive world models (odyssey.ml)

Odyssey has launched Odyssey-2 Pro, a highly advanced world model along with the Odyssey API, marking a pivotal moment for world models akin to OpenAI's GPT-2 for language models. This new world simulator can predict and simulate environments in real-time, leveraging vast amounts of video and interaction data for improved physics, dynamics, and visual fidelity. The Odyssey-2 Pro model can produce continuous interactive video simulations at 720P and 22 FPS, and it supports embedding through three API endpoints: interactive streams, viewable streams, and simulations. Significantly, this release opens the door for developers to explore innovative applications in gaming, education, healthcare, and more, with the potential for unprecedented interactive experiences. With easy integration and the promise of scaling capabilities for performance enhancement, the Odyssey API invites widespread experimentation, much like the initial wave of creativity that followed the release of GPT-2. To encourage this exploration, Odyssey is hosting a hackathon in Palo Alto, further fueling community engagement and development.