Video World Models · 2026-10-28

GameVerse: A Minute-Scale Gameplay Dataset for Long-Horizon Interactive World Modeling

Kang He, Wenshuo Peng, Chuanhao Li, Zihui Gao, Xiaojie Xu, Zhengyuan Lin, Yuzhe Ding, Donghong Ji, Kaipeng Zhang, Yongtao Ge

arXiv:2610.37257PDF

AI-assisted summary based on the paper. Consult the linked source for methods and limitations. Benchmark numbers appear after editor verification.

TL;DR

GameVerse is a minute-scale gameplay video dataset for long-horizon interactive world modeling. The authors report 69.7K clips (1,655 hours, 349.6M frames), including 1,358 hours of continuous two-minute sequences, with temporally stable frame-level camera trajectories and four-level event-aligned captioning (global static, local static, dynamic events, agent actions). They also report that training on GameVerse improves long-horizon temporal consistency and visual quality on a 30-second benchmark.

Why it matters

Interactive world models often drift after roughly one minute, which the authors attribute to insufficient long-horizon training data and coarse captions. GameVerse is designed to supply minute-scale continuous gameplay, validated camera trajectories, and structured, event-aligned semantics, so models can learn temporally coherent state evolution and controllable agent behavior over longer rollouts.

Method

  • Build minute-scale segments from long-form videos by detecting hard cuts (TransNetV2) and gradual semantic transitions (CLIP semantic distance), extracting scene-consistent segments up to 120 seconds.
  • Estimate frame-wise camera poses with ViPE and validate trajectories by filtering three error modes: sudden jumps, high-frequency jitter, and long-term drift; retain clips passing quality checks.
  • Annotate each clip with a four-level hierarchical caption schema (global static, local static, dynamic events, agent actions) aligned to event boundaries, plus multi-stage caption/video quality filtering.
Abstract (from arXiv)

Recent advances in interactive world models have made it possible to generate controllable, dynamic environments from action inputs. Yet current models struggle to maintain coherence beyond roughly one minute of rollout, a limitation rooted not in model capacity but in the scarcity of long-horizon training data. Existing gameplay video datasets typically consist of clips under ten seconds, with captions that describe whole clips in coarse and undifferentiated text, which is insufficient for controllable long-horizon generation. We address this with GameVerse, a minute-scale gameplay video dataset purpose-built for long-horizon interactive world modeling. GameVerse contains 69.7K clips spanning 1,655 hours and 349.6 million frames across diverse first- and third-person gameplay, including 1,358 hours of continuous two-minute sequences, an order of magnitude longer than prior datasets. Clips pass through a unified pipeline of scene-consistent segmentation, quality filtering, camera pose estimation, and event-based caption annotation. Each clip is paired with a temporally stable camera trajectory that remains coherent across minutes, together with captions that separate static scene context from dynamic foreground events and align to event boundaries rather than fixed time windows. Comprehensive analysis shows that GameVerse surpasses existing public datasets in scale, diversity, and annotation quality. Experiments further demonstrate that training on GameVerse substantially extends the temporal horizon of interactive world models.

Related papers

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

preprint

GameVerse: A Minute-Scale Gameplay Dataset for Long-Horizon Interactive World Modeling

Kang He, Wenshuo Peng, Chuanhao Li, Zihui Gao, Xiaojie Xu, Zhengyuan Lin, Yuzhe Ding, Donghong Ji, Kaipeng Zhang, Yongtao Ge

Preprint posted: 2026-10-27 · Original record ↗

arxiv
2610.37257
Explicit source versions (2)
  • v1 · 2026-10-27
  • v2 · 2026-10-29
  1. Preprint available

    Source date: 2026-10-27 · First detected: 2026-10-28 UTC · Historical baseline

    Source evidence (3)
  2. Explicit source version detected · v2

    Source date: 2026-10-29 · First detected: 2026-11-01 UTC · Historical baseline

    Source evidence (1)
Abstract · arxiv

Recent advances in interactive world models have made it possible to generate controllable, dynamic environments from action inputs. Yet current models struggle to maintain coherence beyond roughly one minute of rollout, a limitation rooted not in model capacity but in the scarcity of long-horizon training data. Existing gameplay video datasets typically consist of clips under ten seconds, with captions that describe whole clips in coarse and undifferentiated text, which is insufficient for controllable long-horizon generation. We address this with GameVerse, a minute-scale gameplay video dataset purpose-built for long-horizon interactive world modeling. GameVerse contains 69.7K clips spanning 1,655 hours and 349.6 million frames across diverse first- and third-person gameplay, including 1,358 hours of continuous two-minute sequences, an order of magnitude longer than prior datasets. Clips pass through a unified pipeline of scene-consistent segmentation, quality filtering, camera pose estimation, and event-based caption annotation. Each clip is paired with a temporally stable camera trajectory that remains coherent across minutes, together with captions that separate static scene context from dynamic foreground events and align to event boundaries rather than fixed time windows. Comprehensive analysis shows that GameVerse surpasses existing public datasets in scale, diversity, and annotation quality. Experiments further demonstrate that training on GameVerse substantially extends the temporal horizon of interactive world models.