aiqbench analysis
This exploratory study investigates whether a large multimodal language model can function as an embodied robot controller without prior simulation training or sensorimotor experience, challenging traditional views on the necessity of physical bodies for spatial reasoning. The research demonstrates emergent capabilities in object tracking and action sequencing using only low-resolution visual input, though it also highlights significant perceptual limitations inherent to this approach. These findings offer critical insights into the potential decoupling of world-model development from direct embodiment training within next-generation AI architectures.
Embodied GPT-5.1: Evidence of a World Model?
arXiv:2607.23899v1 Announce Type: cross Abstract: This exploratory study examines whether a large multimodal language model, GPT-5.1, can serve as the high-level controller of a physical mobile robot despite having no prior embodiment, no training in simulated environments, and no exposure to sensorimotor experience. Using only low-resolution first-person images and a discrete action set, the model was tasked with navigation and object-directed behaviors such as locating and contacting a target toy. Across multiple trials, GPT-5.1 demonstrated emergent capabilities that suggest elements of spatial reasoning and physical understanding. These included maintaining short-term memory of object locations after they left the camera frame, inferring the physical consequences of its own movements, and executing coherent action sequences such as colliding with an object and reversing to visually verify the outcome. At the same time, the model displayed inefficiencies and perceptual limitations, including imprecise alignment strategies and occasional misidentification of distant distractors. Overall, the results indicate that GPT-5.1 exhibits signs of world-model-like behavior in an embodied setting, despite the absence of any embodiment-related training, a finding that challenges long-standing views in cognitive science and robotics which hold that a physical body is a necessary prerequisite for developing such forms of intelligence. The findings motivate deeper investigation into the emergence, limits, and robustness of physical understanding in large language models.
Title:Embodied GPT-5.1: Evidence of a World Model?
Abstract:This exploratory study examines whether a large multimodal language model, GPT-5.1, can serve as the high-level controller of a physical mobile robot despite having no prior embodiment, no training in simulated environments, and no exposure to sensorimotor experience. Using only low-resolution first-person images and a discrete action set, the model was tasked with navigation and object-directed behaviors such as locating and contacting a target toy. Across multiple trials, GPT-5.1 demonstrated emergent capabilities that suggest elements of spatial reasoning and physical understanding. These included maintaining short-term memory of object locations after they left the camera frame, inferring the physical consequences of its own movements, and executing coherent action sequences such as colliding with an object and reversing to visually verify the outcome. At the same time, the model displayed inefficiencies and perceptual limitations, including imprecise alignment strategies and occasional misidentification of distant distractors. Overall, the results indicate that GPT-5.1 exhibits signs of world-model-like behavior in an embodied setting, despite the absence of any embodiment-related training, a finding that challenges long-standing views in cognitive science and robotics which hold that a physical body is a necessary prerequisite for developing such forms of intelligence. The findings motivate deeper investigation into the emergence, limits, and robustness of physical understanding in large language models.
6 pages, 16 figures. Published in the 2026 Brazilian Conference on Robotics (CROS)
Robotics (cs.RO); Artificial Intelligence (cs.AI)
arXiv:2607.23899 [cs.RO]
(or arXiv:2607.23899v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2607.23899 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
R. Spinelli and T. C. Martins, "Embodied GPT-5.1: Evidence of a World Model?," 2026 Brazilian Conference on Robotics (CROS), pp. 394-399, 2026
https://doi.org/10.1109/CROS69211.2026.11565684 Focus to learn more DOI(s) linking to related resources
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 357 words
Related coverage
Reasons to believe current AI models are conscious
LessWrong · Jul 17, 2026
[AINews] Reality Checks on AI News (Yegge shuts down Gas Town, Databricks’ +60% Astra cost)
Latent Space · Sep 17, 2026
Rethinking World Models for Safety-Critical Embodied Systems
arXiv · Sep 5, 2026
CST-WM: A Causally Structured World Model for Embodied Visual Tracking
arXiv · Sep 9, 2026