BeingBeyond/Being-H

Being-H is BeingBeyond's family of human-centric embodied foundation models.

What it solves

Being-H provides human-centric embodied foundation models designed to help robots generalize actions across different hardware (cross-embodiment generalization) and reason about future states using egocentric video data.

How it works

The project consists of a family of models with different specializations:

  • Being-H0.7 is a World-Action Model (WAM) that uses egocentric videos to perform future-aware latent reasoning.
  • Being-H0.5 is a Vision-Language-Action (VLA) model that utilizes a unified action space to enable generalization across different robot embodiments.
  • Being-H0 is the initial VLA release focused on pretraining from large-scale human videos.

Who it’s for

It is designed for robotics researchers and developers working on embodied AI, specifically those focusing on dexterous manipulation, cross-embodiment learning, and the use of human video data for robot training.

Highlights

  • Cross-Embodiment Generalization: Enables models to work across different robot types via a unified action space.
  • World-Action Modeling: Integrates latent reasoning about the future into the action process.
  • Human-Centric Learning: Leverages egocentric and large-scale human videos for pretraining.
  • Educational Resources: Includes a dedicated tutorial workspace (Being-H-EDU) for post-training and deployment on robots like the SO101.

Related

  • Project
    X-Square-Robot/wall-x

    An open-source framework for embodied foundation models that enables robots to perform complex physical manipulations by bridging vision-language understanding with action prediction.

  • Project
    Robbyant/lingbot-vla-v2

    LingBot‑VLA 2.0 is an open‑source Vision‑Language‑Action foundation model for robots. It unifies many robot embodiments into a 55‑dim action vector, uses MoE layers and dual‑query distillation (depth + future video) to learn predictive dynamics, and is trained on ~60 k h of robot and egocentric video data. The repo provides conda‑based installation, scripts to download 6 B‑parameter checkpoints, tools for data preparation, post‑training on custom datasets, and evaluation on the RoboTwin benchmark and real robots. Reported results show noticeable gains over the 1.0 version on bimanual and mobile manipulation tasks.

  • Project
    alibaba-damo-academy/RynnBrain

    An embodied foundation model family that bridges vision-language understanding with real-robot control through 3D grounding and contact point prediction.

  • Project
    Open-X-Humanoid/HEX

    HEX is a whole-body vision-language-action framework that enables humanoid robots to perform complex manipulation tasks by aligning data across different robot embodiments.

  • Project
    OpenWAM-Official/OpenWAM

    An open research stack for systematically developing World-Action Models (WAMs) that combine world knowledge and action learning for robotics.