Robbyant is deeply investigating the utility of LingBot-VLA 2.0 in real-world business scenarios
Robbyant has reportedly announced the upgrade and open-source release of LingBot-VLA 2.0, marking a significant move to deliver leaps in morphological generalization, degrees of freedom support and efficient implementation serving as a more advanced universal brain for adaptable robots.
While the embodied AI industry has seen substantial progress in hardware and control systems, and the lack of a unified intelligence remains a primary bottleneck for industrial-scale deployment. It is crucial to note that LingBot-VLA 2.0 addresses this major flaw by dramatically expanding its initial training data and architectural capabilities.
LingBot-VLA was initially trained on 60,000 hours of high-quality real-world physical data. This combined scattered information from 50,000 hours of cleaned real-robot interaction data and 10,000 hours of egocentric action trajectories.
Meanwhile, in terms of cooperative manipulation on Shanghai Jiao Tong University’s GM-100, LingBot-VLA 2.0’s achieved leading project momentum and outperformed both RT-1-0.5 and GR00T N1.7, demonstrating its superior cross-domain adaptability.
Robbyant is actively testing the applications of LingBot-VLA 2.0 in real-world business scenarios. Furthermore it is collaborating with companies including GenRobot.ai to build standardized data ecosystems.
For those unversed, Robbyant is an embodied intelligence company within Ant Group committed to promoting the field through advanced software and hardware technologies.
It is pertinent to mention here that Robbyant independently develops foundational large models for embodied AI and aggressively investigates devices in an effort to create robotic companions.