SOURCE-LINKED INTELLIGENCE

SolarWM: Open Data and Scalable Training for Long-Horizon Video World Models

arXiv · AI, language, vision and robotics · article · Sep 2, 2026 · UTC

We introduce SolarWM, a fully open foundation for building interactive video world models from data preparation through long-horizon inference. Training across heterogeneous data sources and video backbones is challenging: datasets differ in temporal scale, camera geometry, visual quality, motion, and captioning styles, while video generators use distinct representations and architectures. Naive data mixing and model-specific implementations therefore produce inconsistent supervision and make results difficult to reproduce and compare. SolarWM addresses this coupling with a reconfigurable mult

Read original source ↗ Open in workspace

recordType

paper

region

Global