Ant Group’s Robbyant Open-Sources LingBot-VLA to Standardize Robot ‘Brains’

  • Robbyant, an Ant Group subsidiary, open-sourced LingBot-VLA, a vision-language-action (VLA) model designed as a universal brain for robots.
  • The model achieved higher task success rates on the GM-100 benchmark and RoboTwin 2.0 simulation benchmark compared to other models.
  • LingBot-VLA was pre-trained on over 20,000 hours of real-world interaction data across nine mainstream dual-arm robot configurations.
  • The open-source release includes model weights, a production-ready codebase, and tools for data processing, fine-tuning, and automated evaluation.
  • Ant Group launched InclusionAI, a technological ecosystem aimed at advancing AGI through open collaboration.

Ant Group’s open-sourcing of LingBot-VLA addresses a critical bottleneck in robotics: the high cost and low reusability of training models for different hardware. By providing a universal brain that works across multiple robot morphologies, Ant Group is positioning itself as a key player in the embodied AI space. This move aligns with broader industry trends toward open collaboration in AI development, particularly in the pursuit of Artificial General Intelligence (AGI). The release of LingBot-VLA could accelerate the integration of AI into physical applications, from elderly care to medical assistance.

Adoption Pace
How quickly developers will integrate LingBot-VLA into diverse robot platforms and real-world applications.
Performance Scaling
Whether LingBot-VLA can maintain high success rates across increasingly complex and varied environments.
Competitive Response
How rival robotics and AI firms will react to Ant Group’s open-source move and its potential to set industry standards.