EvolvingWorld: An Open-Schema Framework for Co-Evolving Role-Play Agents and World Model in Interactive Literary World

This paper introduces EvolvingWorld, a framework and benchmark for character and world co-evolution in interactive literary worlds. Existing systems either treat interactive literary simulation as static persona imitation or isolated scene generation, failing to capture how characters and worlds evolve together over time. To address this, EvolvingWorld models literary simulation as a long-horizon process where characters interact, scenes progress, and character and world states are persistently updated. Unlike prior systems relying on fixed schemas, EvolvingWorld adopts an open-schema framework to support simulation across diverse literary worlds. The framework consists of two coupled modules: a Character Agent for multicharacter role-play and persistent profile evolution, and an LLM-based World Model for global and location/entity-level state maintenance and scene progression. Based on this architecture, we formulate 7 trainable tasks for scene initialization, interaction generation, and state update. We construct a dataset from 57 books, producing 138,596 supervised training samples and 222 snapshots for testing. Furthermore, we introduce a trajectory-level LLM-as- Judge evaluation protocol spanning 10 dimensions and 20 metrics.

Introduction. Large language models (LLMs) have enabled fluent role-playing agents that imitate fictional characters and sustain persona-grounded dialogue (Shao et al., 2023; Wang et al., 2024, 2025b; Zhou et al., 2024; Xu et al., 2026a,b). Yet simulating a literary world poses a harder long-horizon challenge: as a story unfolds, characters revise beliefs, motivations, and relationships, while locations, objects,

Discussion / Conclusion. We presented EvolvingWorld, a framework and benchmark for simulating interactive, persistently evolving worlds. By coupling a Character Agent with a World Model, maintaining explicit openschema states, and decomposing long-horizon interaction into seven trainable tasks, EvolvingWorld provides a concrete foundation for studying persistent world evolution beyond isolated role-play. Our results show that this design leads to more coherent long-horizon simulations across diverse backbones. We hope EvolvingWorld can serve as a useful step toward richer literary agents, controllable interactive worlds, and long-horizon role-playing systems. EvolvingWorld has three main limitations. First, it models the world as a single objective state shared by all characters, whereas literary characters often perceive and remember the same world differently. For example, one character may view the world as benevolent while another sees it as hostile, and a character may misremember an object as being on a table when it is actually on a chair. Such subjective perceptions and imperfect memories can shape character behavior, but maintaining separate perceived worlds for individual characters would substantially increase system complexity.