| LingBot-VLA 2.0 | Robbyant / Ant Group | China | 2026-07-08 | Open-source VLA foundation model | 6B | 20 robot configurations | Official real-robot application and evaluation material | Open | Configuration dependent | Official GitHub | Robbyant — LingBot-VLA 2.0 |
| XR-1 | Beijing Humanoid Robot Innovation Center | China | 2025-12-18 | Cross-embodiment VLA | Not represented as one comparable public deployment number | Six real robot embodiments in published evaluation | >14,000 reported real-world rollouts across >120 tasks in the paper | Open project artifacts | Not disclosed here | Official project repository / paper | Open X-Humanoid — XR-1 |
| RDT-1B | Tsinghua University / collaborators | China | 2024-10-10 | Diffusion robot foundation model | 1.2B | Single-arm, dual-arm and mobile-manipulator-compatible action spaces | Real-robot bimanual evaluations | Open | Not standardized here | Official project page | RDT-1B official project |
| TurboVLA | Huazhong University of Science and Technology / Huawei | China | 2026-07-29 | Efficient VLA | 216.1M LIBERO configuration; 0.4B bimanual configuration | LIBERO simulation, RoboTwin bimanual setup and AgileX Piper real arm | 160 real-robot trials on AgileX Piper across four tabletop tasks | Open code | 31.2 ms (216.1M configuration, batch 1, RTX 4090) | Official project / author paper | TurboVLA official project |
| ChatVLA | Midea AI | China | 2025 | VLA | Not normalized here | Research manipulation platforms | Research evaluation reported by authors | Open repository and weights | Not disclosed here | Official repository | Midea AI — ChatVLA |
| LingBot-VA | Robbyant / Ant Group | China | 2026 | World-action / video-action model | Not disclosed in one comparable public figure | Simulation and robot-control workflows | Official project demos and deployment instructions | Open repository | Configuration dependent | Official repository | Robbyant — LingBot-VA |