Spatial-IQ: Deconstructing Spatial Intelligence via Hierarchical Capability Tests
cs.CV, cs.AI
Submitted: 2026-07-24
Updated: 2026-11-09
Code: https://github.com/NVIDIA/Spatial-IQ
Project page: https://nvidia.github.io/Spatial-IQ
Terminology
Sources
- Qwen3-VL Technical Report
- HunyuanImage 3.0 Technical Report
- BabyVision: Visual Reasoning Beyond Language
- SpaCE-10: A Comprehensive Benchmark for Multimodal Large Language Models in Compositional Spatial Intelligence
- VLA-0: Building State-of-the-Art VLAs with Zero Modification
- DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning
- EgoExoBench: A Benchmark for First- and Third-person View Video Understanding in MLLMs
- Tulu 3: Pushing Frontiers in Open Language Model Post-Training
- WorldModelBench: Judging Video Generation Models As World Models
- Towards World Simulator: Crafting Physical Commonsense-Based Benchmark for Video Generation
- WorldSimBench: Towards Video Generation Models as World Simulators
- SAT: Dynamic Spatial Aptitude Training for Multimodal Language Models
- DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
- OpenAI GPT-5 System Card
- Kimi K2.5: Visual Agentic Intelligence
- Qwen-Image Technical Report
- SpatialBench: Benchmarking Multimodal Large Language Models for Spatial Cognition
- MMSI-Bench: A Benchmark for Multi-Image Spatial Intelligence
- DAPO: An Open-Source LLM Reinforcement Learning System at Scale
- 3ViewSense: Spatial and Mental Perspective Reasoning from Orthographic Views in Vision-Language Models
Related papers
- TEA: Text Encoder Alignment for Robust Concept Erasure in Text-to-Image Models
- Do Egocentric Video-Language Models Capture Both Hand- and Object-Centric Cues?
- Aura: Consistent Multi-Subject Video Generation via VLM-Grounded Semantic Alignment
- PVRF: All-in-one Adverse Weather Removal via Prior-modulated and Velocity-constrained Rectified Flow
- DLWM: Diverse Latent World Models for Efficient Multimodal Reasoning
- Out of Sight, Still in Mind: Token Compression for Omni-LLMs