Video World Models · 2026-10-28
Hy-Nomas: Towards Free-form Interactive World Generation
Qiaoling Chen, Jiaxiang Cheng, Xueqing Deng, Tiankai Hang, Ji Hou, Jiaqi Li, Yifan Liu, Tao Lu, Junshu Tang, Yu Tang, Yao Teng, Yu Tian, Heng Wang, Weizhi Wang, Xiyao Wang, Xilun Wu, Mingyu Yang, Yangbin Yu, Rui Zhao, Longjun Cao, Linghao Chen, Haoyi Duan, Jiaming Han, Mingji Han, Enhua Jiang, Shunkai Li, Yanzhe Liang, Qinwei Ma, Cong Wang, Lei Wang, Bing Wu, Ke Xing, Ray Ying, Zhentao Yu, Key Zhang, Yuan Zhou, Zhengguang Zhou, Kai Zou
AI-assisted summary based on the paper. Consult the linked source for methods and limitations. Benchmark numbers appear after editor verification.
TL;DR
Hy-Nomas is an 8B-parameter video world model with a diffusion backbone that generates variable-length video segments autoregressively across multiple interaction turns. It follows free-form text instructions and continuous camera control using structured, action-aware causal attention and a vision-language Prompt Enhancer that converts the user request and cached interaction history into conditions for the next segment.
Why it matters
The authors focus on interactive world generation where successive commands must be followed while preserving relevant history and action consequences. Their design separates persistent descriptive context from temporally localized action instructions, and adds explicit continuous camera control across turns, aiming to enable rollouts that reflect both commands and prior interaction history.
Method
- Video world generation via segment-wise autoregressive diffusion: generate one segment at a time conditioned on current prompt, optional camera trajectory, and cached keys/values from preceding segments.
- Action-aware causal attention with structured captions: descriptive context remains accessible across segments while action instructions are restricted to their intended temporal windows; a Prompt Enhancer structures user requests and multimodal histories.
- Multi-stage curriculum for interactive capabilities: progressively introduce segment causality, action-aware control, camera control, training on model-generated histories, then diffusion distillation; release model weights, code, and benchmarks.
Abstract (from arXiv)
Interactive world generation requires models to follow successive commands while preserving relevant history and action consequences. We present Hy-Nomas, a video world model with an 8B diffusion backbone that supports free-form text instructions and continuous camera control across multiple interaction turns. It generates variable-length segments conditioned on current inputs and cached history. Our central design aligns structured video captions with action-aware causal attention: descriptive context remains accessible across segments, while action instructions are restricted to their intended temporal windows. A vision-language Prompt Enhancer (PE) translates user requests and multimodal interaction histories into structured conditions for the next segment. A multi-stage curriculum progressively introduces causal generation, text interaction, camera control, and training on model-generated histories, followed by diffusion distillation. In human evaluation, Hy-Nomas shows superior performance over existing world models. We release model weights, code and benchmarks, together with a lightweight adaptation workflow, to support reproducible research on interactive world generation.
Related papers
- Oneira: From Open-Ended Generation to Open-World Interaction in Video World Models
- FutureWorld: A Plug-and-Play Causal Next-State Predictor for Interactive Video World Exploration
- Reading, Writing, Sharing, and Interacting with the World through Video
- AgiWorld: An Agentic Video World Model for Building Persistent Worlds
- BoxWorld: Let World Engines Control and Video Generators Render
Publication record
Bibliographic evidence describes where and when a record appeared. It does not validate scientific claims.
First public release: 2026-10-27 · Indexed by Pub-AI: 2026-10-28 UTC
Hy-Nomas: Towards Free-form Interactive World Generation
Preprint posted: 2026-10-27 · Original record ↗
- arxiv
- 2610.36729
Explicit source versions (1)
- v1 · 2026-10-27
- Preprint available
Source date: 2026-10-27 · First detected: 2026-10-28 UTC · Historical baseline
Source evidence (2)
- Official evidence ↗ · RSS guid=oai:arXiv.org:2610.36729v1; arxiv:announce_type=new; item pubDate (official metadata)
- Official evidence ↗ · atom:published (official metadata)
Abstract · arxiv
Interactive world generation requires models to follow successive commands while preserving relevant history and action consequences. We present Hy-Nomas, a video world model with an 8B diffusion backbone that supports free-form text instructions and continuous camera control across multiple interaction turns. It generates variable-length segments conditioned on current inputs and cached history. Our central design aligns structured video captions with action-aware causal attention: descriptive context remains accessible across segments, while action instructions are restricted to their intended temporal windows. A vision-language Prompt Enhancer (PE) translates user requests and multimodal interaction histories into structured conditions for the next segment. A multi-stage curriculum progressively introduces causal generation, text interaction, camera control, and training on model-generated histories, followed by diffusion distillation. In human evaluation, Hy-Nomas shows superior performance over existing world models. We release model weights, code and benchmarks, together with a lightweight adaptation workflow, to support reproducible research on interactive world generation.
