H3-World: Turning Language Understanding into World Control
2026-09-01 • Computer Vision and Pattern Recognition
Computer Vision and Pattern RecognitionArtificial Intelligence
AI summaryⓘ
The authors created H3-World, a system that allows detailed control over video scenes generated by a large AI model called MiniMax-H3. They found that using natural language to guide the video generation can be made more precise by linking instructions to specific times in the video. Their method uses attention techniques to keep commands focused on exact moments, improving control without needing extra complicated parts. This approach works well with only a small amount of training and can handle new, unseen situations.
video generationlanguage interfaceworld modeltemporal attentionMiniMax-H3LoRA optimizationzero-shot controlcharacter controlcamera motionsemantic representations
Authors
Danze Chen, Zeqing Wang, Ziyue Lin, Xingyi Yang, Yeying Jin
Abstract
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.