AutoWorldModel-Bench: A State-Centric Benchmark for Automated World-Model Research

AutoWorldModel-Bench：面向自动世界模型研究的以状态为中心的基准测试

World modeling is an unsettled field: architectures, training objectives, and state representations interact in complex ways, and no single recipe dominates across environments. This makes it an ideal testbed for AI coding agents acting as autonomous researchers—a setting in which the improvement direction is not specified in advance, unlike the engineering-to-spec tasks that dominate current agent benchmarks.

世界建模是一个尚未定论的领域：架构、训练目标和状态表示之间存在复杂的相互作用，且目前没有单一的方案能在所有环境中占据主导地位。这使其成为 AI 编程智能体作为自主研究人员进行实验的理想平台——在这种设定下，改进方向并非预先指定，这与当前智能体基准测试中占主导地位的“按规格工程任务”截然不同。

We introduce AutoWorldModel-Bench, a closed-loop benchmark in which frontier coding agents autonomously improve a provided world-model starter under a fixed compute budget. The benchmark spans eight game environments under a unified structured-state representation—ground-truth entity state extracted from each game and consumed through a shared tensor format—which isolates dynamics modeling from perception and enables minutes-per-run iteration.

我们推出了 AutoWorldModel-Bench，这是一个闭环基准测试，前沿编程智能体可以在固定的计算预算下，自主改进所提供的世界模型初始版本。该基准测试涵盖了八个游戏环境，并采用了统一的结构化状态表示——即从每个游戏中提取的真实实体状态，并通过共享的张量格式进行处理——这使得动力学建模与感知过程相隔离，并实现了每轮运行仅需几分钟的迭代速度。

Across 64 sessions, Codex-5.4 and Claude Opus 4.6 improve their starter on 63; in 91% of sessions the winning edit is a non-trivial research-style modification—a new objective, representation, rollout procedure, or architectural change—rather than a hyperparameter tweak. Our benchmark offers a setting in which frontier coding agents can be evaluated on open-ended research rather than engineering-to-spec problems.

在 64 次实验中，Codex-5.4 和 Claude Opus 4.6 在 63 次中成功改进了初始模型；在 91% 的实验中，获胜的编辑方案属于非平凡的研究型修改——例如引入新的目标函数、表示方法、展开过程或架构变更——而非简单的超参数调整。我们的基准测试提供了一个环境，使得前沿编程智能体能够在开放式研究任务中而非仅仅在“按规格工程”问题上得到评估。