Tsinghua Spatial Intelligence Lab (THUSI Lab)

Scaling in Games: Continued Pre-Training for Embodied Agents in Diverse Virtual Worlds. Weights on Hugging Face.

Official implementation of PrismWAM: Decoupling State Representations in World Action Models

Code release for Planning Takes More Than Token Prediction: Causal Plan for Benchmarking and Building Physically Grounded Embodied Reasoners

Video-Arena: A Scalable Agentic Video Benchmark for Spatiotemporal Capability. Videos and GT on Hugging Face.

[ICML 2026] GameVerse: Can Vision-Language Models Learn from Video-based Reflection?

Official project website for DAVIS, a depth-only active-vision framework for humanoid soccer skills (arXiv:2609.28175).

[CVPR 2026] Thinking in 360°: Humanoid Visual Search in the Wild

A Survey on Large Foundation Models as Game Players - Datasets, Models, Harness and Benchmarks

[ECCV 2026] Towards Spatial Supersensing in the Wild