Robbyant Releases LingBot-VLA 2.0: An Open-Source 6B Vision-Language-Action (VLA) Model for Cross-Embodiment Robot Manipulation
Robbyant has open‑sourced LingBot‑VLA 2.0, a 6‑billion‑parameter vision‑language‑action model trained on 60,000 hours of robot and human data. It maps diverse robot embodiments to a unified 55‑dimensional action space.
- LingBot‑VLA 2.0 is a 6 B open‑source VLA model trained on massive robot and human video data.
- It unifies diverse robot embodiments into a single 55‑dimensional action space.
- The architecture employs a token‑level MoE action expert without a load‑balancing loss.
- Open‑sourcing the model lowers entry barriers for robotics developers and researchers.
Robbyant, the robotics AI unit of Ant Group, released LingBot‑VLA 2.0 under the Apache‑2.0 license. The model contains 6 billion parameters and is designed to interpret visual inputs, language commands, and produce robot actions across multiple embodiments.
The training corpus comprises roughly 60 000 hours of data, including 50 000 hours of robot trajectories from 20 different robot configurations and 10 000 hours of egocentric human video. All embodiments are projected into a 55‑dimensional canonical action space, covering arms, dexterous hands, waists, heads and mobile bases.
Architecturally, LingBot‑VLA 2.0 uses a token‑level Mixture‑of‑Experts (MoE) action expert that scales capacity without requiring a load‑balancing loss, simplifying training. The model is positioned as a foundation for developers building cross‑embodiment manipulation systems.
By releasing the model openly, Robbyant aims to accelerate research and commercial applications in robot manipulation, lowering the barrier for teams to experiment with unified action representations.
Provides a ready‑to‑use foundation model for building cross‑embodiment robot control systems.
Enables faster prototyping of robot solutions across different hardware platforms.
Signals growing investment in open‑source robotics AI, potentially expanding market opportunities.
Offers a publicly available research model for studying vision‑language‑action integration.
Advances the ability of robots to understand and act on visual and language cues across varied forms.
- Vision‑Language‑Action (VLA)
- A model type that processes visual inputs and language commands to generate robot actions.
- Mixture‑of‑Experts (MoE)
- A neural architecture that routes tokens to specialized expert sub‑networks to increase capacity efficiently.
- Canonical action space
- A unified representation of robot actions that can be shared across different robot morphologies.
RoboticsHumanoid robots controlled by surgeons did world-first operation on live pigs
RoboticsThe 1X Neo Robot Has Freaky Fast Fingers
RoboticsRobbyant Releases LingBot-VLA 2.0: An Open-Source 6B Vision-Language-Action (VLA) Model for Cross-Embodiment Robot Manipulation
RoboticsThis startup thinks robotics is about to have its ChatGPT moment
RoboticsMistral enters robotics with Robostral Navigate, an 8B model that steers robots using just one camera
AI ToolsHow We Test an AI Product Without Burning Credit
A developer shares practical strategies to validate AI features without exhausting API credits, focusing on mocking and lightweight testing.
BusinessRed Hat will support your RHEL forever now - for a price
Red Hat now offers a paid Long-Life Add-On to extend support for specific RHEL releases indefinitely, ensuring long-term stability for enterprises.
BusinessApple Is Suing OpenAI for Allegedly Stealing Hardware Secrets
Apple has filed a lawsuit against OpenAI, accusing the company of encouraging poached employees to steal confidential hardware designs and supplier details.
AI in battlefield intelligence: Expanding the speed of decision-making - Federal News Network
The U.S. military is integrating AI tools to speed up battlefield intelligence and decision-making processes, according to Federal News Network.
Apple Sues OpenAI, Alleging It Stole Trade Secrets - WSJ
Apple has filed a lawsuit against OpenAI, accusing the company of stealing trade secrets related to its AI technology.
BusinessApple sues OpenAI over alleged trade secret theft
Apple has filed a lawsuit against OpenAI, accusing the company of stealing trade secrets with involvement from its senior leadership and a former employee.