aiqbench analysis
This survey addresses the "embodiment gap" in robot foundation models by distinguishing between reusable semantic representations and specific adaptations required for physical execution on distinct hardware bodies. It proposes a new reporting framework that moves beyond simple success rates to explicitly quantify the remaining work needed when deploying generalized policies onto target robots with unique kinematics or sensors. This analysis is critical for researchers aiming to bridge the divide between scalable AI models and practical, real-world robotic deployment across diverse platforms.
The Embodiment Gap in Robot Foundation Models
arXiv:2608.18433v1 Announce Type: new Abstract: Robot foundation models (RFMs), including vision-language-action (VLA) policies, are often discussed through a scaling view: more data, larger models, and broader benchmarks should improve generalization. In robotics, however, a model can generalize while work still remains before it can run on a robot with a particular body. The work required differs across methods and target robots, and those differences affect practical deployment. We call the gap between reusable models, representations, or data and their use in execution on the target robot the embodiment gap. This survey examines what can be reused across robot embodiments and what must still be implemented on a new robot. We place existing methods on a two-axis map that shows the type of shared structure and the stage at which adaptation is needed for execution on the target robot. We then examine recent work through three overlapping research directions: sharing semantics and perception, sharing robot data and interfaces, and learning correspondence across embodiments. We also propose a reporting framework for adaptation work that success rate alone does not reveal. The framework identifies the work that should be checked when comparing cross-embodiment learning and highlights work that remains on a new robot and questions for future study.
Abstract:Robot foundation models (RFMs), including vision-language-action (VLA) policies, are often discussed through a scaling view: more data, larger models, and broader benchmarks should improve generalization. In robotics, however, a model can generalize while work still remains before it can run on a robot with a particular body. The work required differs across methods and target robots, and those differences affect practical deployment. We call the gap between reusable models, representations, or data and their use in execution on the target robot the embodiment gap. This survey examines what can be reused across robot embodiments and what must still be implemented on a new robot. We place existing methods on a two-axis map that shows the type of shared structure and the stage at which adaptation is needed for execution on the target robot. We then examine recent work through three overlapping research directions: sharing semantics and perception, sharing robot data and interfaces, and learning correspondence across embodiments. We also propose a reporting framework for adaptation work that success rate alone does not reveal. The framework identifies the work that should be checked when comparing cross-embodiment learning and highlights work that remains on a new robot and questions for future study.
32 pages, 4 figures. Published in Transactions on Machine Learning Research (TMLR), August 2026
Robotics (cs.RO); Machine Learning (cs.LG)
arXiv:2608.18433 [cs.RO]
(or arXiv:2608.18433v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2608.18433 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
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 · 310 words
Related coverage
Building a Foundation Stack for General-Purpose Robots
IEEE Spectrum · Jul 13, 2026
Open-H-Embodiment: A Large-Scale Dataset for Enabling Foundation Models in Medical Robotics
arXiv · Jun 6, 2026
Running MolmoAct2 Robot Policy on an AMD Ryzen AI Max+ 395 Mini PC
AMD · Sep 30, 2026
Data Pyramid for Embodied Manipulation
arXiv · Jul 28, 2026