Has GPT-5 Achieved Spatial Intelligence?
GPT-5 sets SoTA but not human‑level spatial intelligence.
My notes below:
This report introduces a unified view of spatial intelligence (SI) for multimodal models and evaluates GPT‑5 and strong baselines across eight fresh SI benchmarks.
GPT‑5 leads overall but is still short of human skill, especially on mentally reconstructing shapes, changing viewpoints, and deformation/assembly tasks.
GPT‑5 leads overall but is still short of human skill, especially on mentally reconstructing shapes, changing viewpoints, and deformation/assembly tasks.
Unified SI schema and fair eval setup
The authors consolidate prior work into six core SI capabilities (Metric Measurement, Mental Reconstruction, Spatial Relations, Perspective‑taking, Deformation & Assembly, Comprehensive Reasoning) and standardize prompts, answer extraction, and metrics to reduce evaluation variance across datasets.
The authors consolidate prior work into six core SI capabilities (Metric Measurement, Mental Reconstruction, Spatial Relations, Perspective‑taking, Deformation & Assembly, Comprehensive Reasoning) and standardize prompts, answer extraction, and metrics to reduce evaluation variance across datasets.
Broad benchmark sweep, heavy compute
Eight recent benchmarks (e.g., VSI‑Bench, SITE, MMSI, OmniSpatial, MindCube, STARE, CoreCognition, SpatialViz) are used with unified protocols; results reflect >1B tokens of evaluation traffic.
Eight recent benchmarks (e.g., VSI‑Bench, SITE, MMSI, OmniSpatial, MindCube, STARE, CoreCognition, SpatialViz) are used with unified protocols; results reflect >1B tokens of evaluation traffic.
GPT‑5 sets SOTA but not human‑level SI.
GPT‑5 tops aggregate scores and sometimes reaches human parity on Metric Measurement and Spatial Relations, yet shows significant gaps on Mental Reconstruction, Perspective‑taking, Deformation & Assembly, and multi‑stage Comprehensive Reasoning.
GPT‑5 tops aggregate scores and sometimes reaches human parity on Metric Measurement and Spatial Relations, yet shows significant gaps on Mental Reconstruction, Perspective‑taking, Deformation & Assembly, and multi‑stage Comprehensive Reasoning.
Thinking Modes
"In High mode, 28 questions exceeded the 15-minute time limit or hit token limit, and were counted as incorrect, resulting in an accuracy of 52.54%; excluding these cases yields 68.89%."
High could lead to better results, but I understand the token limit issues.
"In High mode, 28 questions exceeded the 15-minute time limit or hit token limit, and were counted as incorrect, resulting in an accuracy of 52.54%; excluding these cases yields 68.89%."
High could lead to better results, but I understand the token limit issues.
Hard SI narrows the closed vs open gap
While proprietary models win on average, their advantage evaporates on the hardest SI categories; several open‑source systems perform similarly, far from human ability on MR/PT/DA/CR.
Non‑SI portions (e.g., CoreCognition’s Formal Operation) can reach near‑human levels.
While proprietary models win on average, their advantage evaporates on the hardest SI categories; several open‑source systems perform similarly, far from human ability on MR/PT/DA/CR.
Non‑SI portions (e.g., CoreCognition’s Formal Operation) can reach near‑human levels.
Qualitative analysis exposes failure modes
Case studies show prompt sensitivity for novel‑view generation, blind spots with perspective effects and size constancy, persistent failures on paper‑folding/assembly, and difficulty inferring occluded objects during counting.
Paper: arxiv.org/abs/2508.13142
Case studies show prompt sensitivity for novel‑view generation, blind spots with perspective effects and size constancy, persistent failures on paper‑folding/assembly, and difficulty inferring occluded objects during counting.
Paper: arxiv.org/abs/2508.13142
Generated by Thread Navigator
Press ⌘ + S to quick-export