The Year in AI Papers
The year/Independent research

Paper 2604.07430

HY-Embodied-0.5: Embodied Foundation Models for Real-World Agents

Tencent Robotics X, HY Vision Team, :, Xumin Yu, Zuyan Liu, Ziyi Wang, He Zhang, Yongming Rao, Fangfu Liu, Yani Zhang, Ruowen Zhao, Oran Wang, et al.

Published
Apr 2026
Research lab
Independent
Citations
12
GitHub
843 stars

01 In brief

Summary

Tencent's HY-Embodied-0.5 is a family of vision-language foundation models designed for real-world embodied agents, bridging the gap between general VLMs and physical-world tasks.

The suite includes an efficient 2B-activated-parameter model (MoT-2B) for edge deployment and a powerful 32B-activated-parameter model (MoE-A32B) for complex reasoning.

Key innovations include a Mixture-of-Transformers (MoT) architecture with modality-specific computing, visual latent tokens, a native-resolution ViT encoder, and an iterative self-evolving post-training paradigm with reinforcement learning and on-policy distillation.

Evaluated on 22 benchmarks covering visual perception, spatial reasoning, and embodied understanding, the MoT-2B model outperforms similarly sized state-of-the-art models on 16 benchmarks, while the 32B variant achieves an average score of 67.0%, surpassing Gemini 3.0 Pro (63.6%).

In real-world robot control tasks, a VLA model built on the MoT-2B foundation achieves success rates of 85% (packing), 80% (stacking), and 75% (hanging), outperforming π0 and π0.5 baselines on the most challenging task.

Code and models are open-sourced at https://github.com/Tencent-Hunyuan/HY-Embodied.

02 From the paper

Abstract

We introduce HY-Embodied-0.5, a family of foundation models specifically designed for real-world embodied agents. To bridge the gap between general Vision-Language Models (VLMs) and the demands of embodied agents, our models are developed to enhance the core capabilities required by embodied intelligence: spatial and temporal visual perception, alongside advanced embodied reasoning for prediction, interaction, and planning. The HY-Embodied-0.5 suite comprises two primary variants: an efficient model with 2B activated parameters designed for edge deployment, and a powerful model with 32B activated parameters targeted for complex reasoning. To support the fine-grained visual perception essential for embodied tasks, we adopt a Mixture-of-Transformers (MoT) architecture to enable modality-specific computing. By incorporating latent tokens, this design effectively enhances the perceptual representation of the models. To improve reasoning capabilities, we introduce an iterative, self-evolving post-training paradigm. Furthermore, we employ on-policy distillation to transfer the advanced capabilities of the large model to the smaller variant, thereby maximizing the performance potential of the compact model. Extensive evaluations across 22 benchmarks, spanning visual perception, spatial reasoning, and embodied understanding, demonstrate the effectiveness of our approach. Our MoT-2B model outperforms similarly sized state-of-the-art models on 16 benchmarks, while the 32B variant achieves performance comparable to frontier models such as Gemini 3.0 Pro. In downstream robot control experiments, we leverage our robust VLM foundation to train an effective Vision-Language-Action (VLA) model, achieving compelling results in real-world physical evaluations. Code and models are open-sourced at https://github.com/Tencent-Hunyuan/HY-Embodied.