Ant Group’s Robbyant Releases Open-Source LingBot-VLA 2.0 for Embodied AI
Event summary
- Robbyant, an Ant Group subsidiary, released LingBot-VLA 2.0 on July 8, 2026, upgrading its vision-language-action model for embodied AI.
- The model was pre-trained on 60,000 hours of real-world physical data from 20 distinct robot morphologies across 17 manufacturers.
- LingBot-VLA 2.0 achieved leading performance in dual-arm manipulation and long-horizon mobile manipulation tasks, outperforming competitors like π0.5 and GR00T N1.7.
- The model is optimized for efficient post-training with latency under 130 milliseconds on an RTX 4090, lowering commercial deployment costs.
- Robbyant has partnered with hardware and enterprise customers for pilot testing in retail sorting, logistics, and industrial automation.
The big picture
The release of LingBot-VLA 2.0 underscores Ant Group’s push to dominate the embodied AI space, addressing a critical bottleneck in industrial-scale robotics deployment. By open-sourcing the model and partnering with leading hardware manufacturers, Robbyant aims to accelerate real-world applications in retail, logistics, and automation. The strategic move positions Ant Group as a key player in shaping the future of intelligent robotic systems.
What we're watching
- Adoption Pace
- The pace at which LingBot-VLA 2.0 will be integrated into commercial applications and its impact on Ant Group’s broader AI strategy.
- Competitive Edge
- Whether Robbyant can sustain its performance lead in embodied AI against competitors like π0.5 and GR00T N1.7.
- Open-Source Impact
- How the open-sourcing of LingBot-VLA 2.0 will influence industry collaboration and innovation in robotics.
Related topics