Ant Robbyant Open-Sources LingBot-VLA 2.0: 60k Hours Pre-training, 20 Robot Configs
CyberRobooo · x · 2026-07-08
Ant Group's Robbyant team open-sourced LingBot-VLA 2.0, a vision-language-action foundation model pre-trained on 60,000 hours of data (including 10,000 hours of first-person human operation videos). It uniformly supports coordinated control across 20 robot configurations—including Astribot, AGIBOT, and X-humanoid—covering the head, waist, dexterous hands, and mobile base. The new version shows significant improvements in long-horizon mobile manipulation tasks (like organizing a fridge or cleaning a stove, requiring movement plus dual-arm and hand coordination) and surpasses π₀.₅ and GR00T N1.7 on the GM-100 bimanual manipulation benchmark.
Related event: Ant Lingbo Open-Sources LingBot-VLA 2.0 for Multi-Robot Generalization(18 posts)→
More from Embodied
- WanGP v18.81 adds Tencent's Hy-Nomas world model, playable on 8GB VRAM — cocktailpeanut · 2026-10-27
- NVIDIA open-sources FLORA, a robotics data engine ~125x cheaper than GPT-6 Astra — nvidia · 2026-10-27
- Embodied AI startups need a different operational discipline, investor says — ___Mufasaa · 2026-10-27
- π0.5 study: raising demos from 10 to 438 lifts Place Ball in Bowl success from 7% to 83% — stingning · 2026-10-27
- GPT-6 Astra controls dexterous hands zero-shot at just 5.2% success in 500 trials — stingning · 2026-10-27
- HandsOn benchmark: 66 verified tasks and 18,000 human demos for dexterous manipulation — stingning · 2026-10-27