H3-World：将语言理解转化为世界控制

摘要

我们提出 H3-World，一个高效的框架，可将 33B 参数的 MiniMax-H3 视频生成器转化为交互式世界模型。我们的关键发现是：随着大型视频生成器能力的增强，语言正成为一种自然的控制接口。例如，MiniMax-H3 已支持通过自然语言指令对角色行为和摄像机运动进行零样本控制。在此基础上，H3-World 将该粗粒度语言接口转化为精确且时间锚定的世界控制，而无需引入专门的动作模块。具体而言，我们将每个动作表示为角色指令与摄像机指令的结构化组合，并将其与相应时间区间的视频潜在表示对齐。为了实现时间上的精确控制，我们进一步引入时间注意力路由机制，将每条指令限制在其预期的时间区间内，从而减少跨动作的控制泄漏。重要的是，H3-World 直接复用大规模视频预训练阶段学习到的语义表示，仅需轻量级适配。仅使用 8,000 个游戏样本、10,000 步 LoRA 优化和 0.199% 的可训练参数，H3-World 便实现了有效的角色与摄像机控制，同时保持了优异的生成质量，并能泛化到未见过的场景。这些结果表明，大型视频生成器中涌现出的控制能力可以被高效地转化为交互式世界控制。

English

We present H3-World, an efficient framework that turns the 33B MiniMax-H3 video generator into an interactive world model. Our key finding is that, as large video generators become more capable, language is emerging as a natural interface for control. MiniMax-H3, for example, already supports zero-shot control of character behavior and camera motion through natural-language instructions. Building on this, H3-World turns this coarse language interface into precise, temporally grounded world control, without introducing dedicated action modules. Specifically, we represent each action as a structured combination of character and camera instructions, and align them with the corresponding temporal video latents. To make the control temporally precise, we further introduce temporal attention routing, which restricts each instruction to its intended time interval and reduces control leakage across actions. Importantly, H3-World directly reuses the semantic representations learned during large-scale video pretraining and requires only lightweight adaptation. With only 8,000 gameplay samples, 10,000 LoRA optimization steps, and 0.199% trainable parameters, H3-World achieves effective character and camera control while preserving strong generation quality. It also generalizes to unseen scenarios. These results show that the control capabilities emerging in large video generators can be efficiently transformed into interactive world control.