From Pixels to States: Rethinking Interactive World Models as Game Engines

2026-07-15

AI Summary

Existing video generation models struggle to meet the core requirements of interactive game worlds—namely, rule-driven behavior, state persistence, and real-time responsiveness. This work proposes a novel modeling paradigm centered on explicit game states, decomposing interactive world modeling into four key dimensions: action control, state evolution, state-observation consistency, and real-time generation. To support this framework, the authors construct a large-scale, frame-aligned data engine tailored for *Black Myth: Wukong*, integrating native game engine mechanics, structured semantic annotations, and a scalable data collection pipeline. They release the first high-quality dataset featuring ground-truth game states, actions, and semantic labels, spanning over 90 hours of gameplay. This resource establishes a benchmark for state-aware generative AI, advancing its evolution toward rule-consistent and temporally coherent game engines.

Abstract

Building interactive worlds that respond coherently to player actions has long been a shared goal of computer graphics, games, and artificial intelligence. Recent video generative models provide a data-driven route toward this goal by predicting future observations conditioned on user actions, and are increasingly regarded as potential next-generation game engines. Realizing a genuinely interactive game world, however, requires interaction outcomes that follow rules over evolving game conditions, consequences that persist over long horizons, and a generation loop that operates in real time. Conventional game engines realize these properties through a recurrent action-state-observation loop, in which player actions update an explicit game state according to predefined rules and observations are rendered from the resulting state. Taking this loop as an organizing lens, this paper examines interactive game world modeling along four dimensions: player action control, game state dynamics, state-observation persistence, and real-time interactive generation. For each dimension, we start from the capabilities required by an interactive game world, group existing approaches into representative families, and discuss the strengths and trade-offs of each family. Complementing this analysis, we present a scalable data engine for Black Myth: Wukong that collects over 90 hours of gameplay with frame-aligned player actions, ground-truth game states, and visual observations, together with structured and semantic annotations, as a resource for state-aware game world modeling. We hope this paper offers a clear picture of where the field stands and fosters progress toward interactive game worlds.

Problem

• interactive world models
• game engines
• player actions
• game state dynamics
• real-time generation

Innovation

• interactive world models
• game state dynamics
• real-time generation
• state-aware modeling
• data-driven game engines