Openflows

Signal

2026 年，这些开源的具身大脑模型正在改变着技术格局......-CSDN 博客 · brave · 2026-03-24 Xiaomi releases Xiaomi-Robotics-0, a 4.7-billion parameter vision-language-action (VLA) model featuring a dual-brain architecture where a visual-language brain parses instructions and a motor control brain generates continuous action vectors. The model achieves 80ms latency and 30Hz real-time control on consumer-grade GPUs, claiming record performance across three benchmarks. It targets home service and industrial sorting applications, aiming to reduce hardware costs by two orders of magnitude.

Context

Embodied AI infrastructure is shifting from closed, high-cost systems toward open-weight models deployable on consumer hardware. The dual-brain architecture separates cognitive reasoning from low-latency motor execution, addressing the latency constraints typical of end-to-end VLA models. This release signals a move toward localized inference for physical agents, aligning with the broader Openflows baseline for local model deployment.

Relevance

Xiaomi-Robotics-0 lowers the barrier for entry in robotics development by removing reliance on proprietary cloud APIs. Its open-source nature allows for community-driven optimization and integration with existing agent orchestration frameworks. The model's efficiency on consumer hardware supports the decentralization of physical AI operations, reducing dependency on centralized data centers.

Current State

The model is available as an open-source release with claimed support for 30Hz control loops. It is optimized for NVIDIA consumer GPUs and supports inference on local hardware without high-level abstraction layers. Documentation and weights are being distributed through Xiaomi's open-source channels, with integration plans for home and industrial hardware ecosystems.

Open Questions

Long-horizon task reliability and safety mechanisms for physical interaction in unstructured environments remain to be validated. Integration with existing multi-agent orchestration systems (e.g., OpenClaw, CrewAI) requires further technical specification. The model's performance under varying lighting and environmental conditions outside benchmark settings needs independent verification.

Connections

Xiaomi-Robotics-0 operates within the same infrastructure layer as rynnbrain, providing an alternative open embodied foundation model for grounded robot planning. It interfaces with frameworks like dimensionalos which manage skills-based ROS2 architectures for robot control primitives. The release contributes to the distributed-physical-agent-infrastructure circuit by enabling software-native control over distributed physical systems with reduced latency.