WorldCast: Distributed Multiplayer World Models

2026-10-08

AI Summary

This study addresses the prohibitive computational costs and viewpoint consistency challenges inherent in joint generation for multiplayer world models as player counts scale. To overcome these limitations, this work proposes a decentralized distributed architecture wherein each client independently executes a video generator and a state model. Cross-client synchronization is achieved through camera-aligned player state fields and a shared scene state mechanism, thereby eliminating centralized computational bottlenecks. Evaluated within Counter-Strike 2 (CS2), the proposed method achieves real-time multi-view consistent generation while improving rendering efficiency by over an order of magnitude. Furthermore, it maintains stable visual quality during long-horizon rollouts and demonstrates superior scalability.

Abstract

Multiplayer world models must generate independently controlled views with consistent representations of both players and their shared environment. Most existing approaches coordinate multiple players through joint multi-view generation, whose cost grows with each additional player. We present WorldCast, a distributed multiplayer world model in which each player runs a local client comprising a video generator and a state model. Using recorded player positions and map geometry during training, the state model estimates the player's position from generated video and control inputs. Clients exchange player states and project them into camera-aligned player state fields that guide where and how other players are rendered. Shared scene state enables clients to reuse one another's generated observations to maintain consistent scene appearance across views. Experiments on Counter-Strike 2 demonstrate WorldCast's consistency, real-time performance, and distributed scalability. The camera-aligned player state field improves player rendering rates by over an order of magnitude over joint-generation methods, while shared scene state improves visual consistency over whole rounds. Each client runs in real time and exchanges only player and scene states, enabling scalable multiplayer generation without a centralized computational bottleneck. Image quality remains stable over hour-long rollouts.

Problem

• multiplayer world models
• distributed generation
• multi-view consistency
• scalability

Innovation

• Distributed World Model
• Multiplayer Generation
• Camera-Aligned Player State Field
• Shared Scene State
• Scalable Video Generation