Robbyant Upgrades and Open-Sources LingBot-VLA 2.0 as a Next-Generation Universal Brain for Embodied AI | RoboticsTomorrow
Robbyant, an embodied AI company under Ant Group, today announced the upgrade and open-source release of LingBot-VLA 2.0. Building on its predecessor from January 2026, this next-generation vision-language-action (VLA) model delivers major advances in morphological generalization, degrees of freedom (DoF) support, and deployment efficiency—moving closer to a true "universal brain" for scalable real-world robotics.
While hardware and control systems in embodied AI have evolved rapidly, the lack of a universally adaptable brain remains a key barrier to industrial-scale deployment. LingBot-VLA 2.0 directly tackles this gap with expanded pre-training data and architectural innovations.
Data and Morphological Breadth
The model was pre-trained on 60,000 hours of high-quality real-world physical data, consisting of 50,000 hours of cleaned real-robot interaction data and 10,000 hours of distilled first-person human manipulation data. This diverse dataset spans 20 distinct robot morphologies from 17 manufacturers—including Leju, AgiBot, Unitree, AgileX, Galaxea, Galbot, Astribot, RealMan, Franka, ARX, X-Humanoid, Fourier, MagicLab, Spirit AI, Zerith, Flexiv, and Qinglong. Configurations covered range from single-arm and dual-arm to bipedal and wheeled designs.
Expanded Degrees of Freedom and Whole-Body Control
LingBot-VLA 2.0 extends its operational capabilities to include head, waist, end-effectors (hands), and mobile chassis, enabling highly coordinated whole-body control. This expanded DoF support allows the model to handle more complex, multi-joint tasks in real environments.
Benchmark Performance
On Shanghai Jiao Tong University's GM-100 benchmark for dual-arm manipulation, LingBot-VLA 2.0 achieved leading average task progress scores and success rates on AgileX Cobot Magic and Galaxea R1 Pro platforms. It outperformed both 0.5 and GR00T N1.7, demonstrating superior cross-morphology generalization.
In long-horizon mobile manipulation tasks tested on the ARX Arm + AgileX Chassis and Astribot S1 platforms, the model surpassed a 0.5 score in both task progress and success rates. Its robust cross-domain performance highlights advanced long-sequence task execution and mobile manipulation generalization.
Deployment Efficiency and Cost Reduction
To lower the high costs of post-training and deployment, LingBot-VLA 2.0 introduces a version optimized for efficient fine-tuning. Latency is strictly maintained under 130 milliseconds on an RTX 4090, significantly lowering the barrier for commercial adoption.
Real-World Commercial Pilots
Robbyant is actively piloting LingBot-VLA 2.0 in business scenarios with hardware partners like Leju and Ti5 Robot, and enterprise customers including GuoDa Drugstore and Longsheng Technology. Use cases span retail sorting, logistics, and industrial automation. Additionally, Robbyant is collaborating with companies such as GenRobot.ai to build standardized data ecosystems.
Open-Source Release and Community Engagement
LingBot-VLA 2.0 is fully open-sourced today. Resources are available on:
- GitHub: https://github.com/Robbyant/lingbot-vla-v2
- Hugging Face: https://huggingface.co/collections/robbyant/lingbot-vla-v2
Looking ahead, Robbyant will host developer meetups and release specialized technical toolkits tailored for the developer community.
About Robbyant
Robbyant, part of Ant Group, is dedicated to advancing embodied intelligence through cutting-edge software and hardware. The company develops foundational large models for embodied AI and explores next-generation intelligent devices, aiming to create robotic companions and caregivers that enhance everyday life. Key application areas include elderly care, medical assistance, and household tasks.
For more information, visit www.robbyant.com.
The source for this article is https://www.roboticstomorrow.com/news/2026/07/08/robbyant-upgrades-and-open-sources-lingbot-vla-20-as-a-next-generation-universal-brain-for-embodied-ai/26819/.