RLDX-1 - 2026-05-07 - Research Publication - RLDX-1 is a multimodal robot foundation model for dexterous manipulation. Its GR1 fine-tuned checkpoint is evaluated on the simulated 24-task GR-1 Tabletop benchmark, with 58.7% average success; that checkpoint result does… - source: huggingface.co
UniT: Toward a Unified Physical Language for Human-to-Humanoid Policy Learning and World Modeling - 2026-04-21 - Research Publication - UniT learns shared visual-action tokens for human-to-humanoid policy transfer and action-conditioned world modeling. It evaluates RoboCasa GR1 simulation and physical IRON-R01-1.11 pick-and-place and bimanual pouring… - source: arxiv.org
Modality-Augmented Fine-Tuning of Foundation Robot Policies for Cross-Embodiment Manipulation on GR1 and G1 - 2025-12-01 - Research Publication - This simulation study augments GR00T policies with contact and depth signals across Fourier GR1 and Unitree G1 embodiments. It uses RoboSuite/RoboCasa rollouts and releases simulated G1 apple-to-bowl demonstrations with… - source: arxiv.org
NVIDIA publishes Isaac GR00T N1 open foundation model for humanoid robots - 2025-03-17 - Research Publication - GR00T N1 combines a vision-language model with a diffusion action module and learns from real-robot, human-video and synthetic trajectories. NVIDIA evaluates it in simulation and on physical Fourier GR-1 tabletop… - source: research.nvidia.com