Has GPT-5 Achieved Spatial Intelligence? An Empirical Study

2025-08-18

AI Summary

This study systematically evaluates the spatial intelligence of GPT-5 and leading open- and closed-source multimodal models, aiming to delineate their capabilities and limitations in spatial understanding and reasoning. We introduce the first unified taxonomy of spatial tasks—spanning eight standardized benchmarks—and identify critical weaknesses, including geometric reasoning and relative position judgment. Through large-scale empirical evaluation (>1B tokens), integrating quantitative benchmarking with cross-scenario qualitative analysis, we find that closed-source models—including GPT-5—do not significantly outperform state-of-the-art open-source models on the most challenging tasks, and all models remain substantially below human-level performance. Our key contributions are threefold: (1) a scalable, principled spatial task taxonomy; (2) exposure of systematic biases in existing spatial evaluation protocols; and (3) the first empirical demonstration that current multimodal models consistently fail on spatially intuitive scenarios solvable by humans.

Technology Category

• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Machine Learning: Large Multimodal Models (LMMs)
• Computer Vision: Multi-modal Vision

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Web search models and ranking
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Multi-modal models have achieved remarkable progress in recent years. Nevertheless, they continue to exhibit notable limitations in spatial understanding and reasoning, which are fundamental capabilities to achieving artificial general intelligence. With the recent release of GPT-5, allegedly the most powerful AI model to date, it is timely to examine where the leading models stand on the path toward spatial intelligence. First, we propose a comprehensive taxonomy of spatial tasks that unifies existing benchmarks and discuss the challenges in ensuring fair evaluation. We then evaluate state-of-the-art proprietary and open-source models on eight key benchmarks, at a cost exceeding one billion total tokens. Our empirical study reveals that (1) GPT-5 demonstrates unprecedented strength in spatial intelligence, yet (2) still falls short of human performance across a broad spectrum of tasks. Moreover, we (3) identify the more challenging spatial intelligence problems for multi-modal models, and (4) proprietary models do not exhibit a decisive advantage when facing the most difficult problems. In addition, we conduct a qualitative evaluation across a diverse set of scenarios that are intuitive for humans yet fail even the most advanced multi-modal models.

Problem

• Evaluating GPT-5's spatial intelligence capabilities
• Assessing limitations in multi-modal models' spatial reasoning
• Comparing AI and human performance on spatial tasks

Innovation

• Proposed comprehensive taxonomy for spatial tasks
• Evaluated models on eight benchmarks extensively
• Identified challenging spatial problems for models