✦
Summary
LingBot-World 2.0 is Robbyant interactive world-model project for generating environments that respond to navigation, actions, and high-level direction. Its LingBot-World-Infinity demonstrations emphasize extended exploration, shared control, and evolving scenes rather than fixed-duration generated clips.
Key Features
The model supports interactive exploration with immediate visual feedback.
The project reports real-time 720p generation at 60 frames per second.
Player and director roles can collaboratively steer a shared generated environment.
Commands and scene-level interventions can guide world evolution.
Demonstrations emphasize persistent exploration over extended generation horizons.
Action-conditioned dynamics are presented for embodied simulation and future-state prediction.
The repository releases 14B causal-fast, causal-pretrained, bidirectional, and 1.3B causal-fast variants.
The authors identify long-term memory, extended interaction, and physical understanding as open challenges.
About LingBot-World 2.0
The model learns action-conditioned visual dynamics from egocentric, synthetic, and web video. The project reports real-time generation at 720p and 60 frames per second with sub-second control latency. Player and director inputs can jointly influence the generated world, including subjects, events, and environmental changes.
The work supports research in interactive entertainment, embodied simulation, and future-state prediction. The repository releases 14B and 1.3B variants with inference code under CC BY-NC-SA 4.0, superseding the landing page release notice. Long-term memory and faithful physical reasoning remain research challenges, and the internal deployment stack is not released.