Infinigence open-sources APXInf: cuts embodied AI on-robot inference latency from 278ms to 26ms
青稞AI · wechat · 2026-09-15
Infinigence (Wuxin Qiong), with Tsinghua and Shanghai Jiao Tong University, open-sourced APXInf, an edge inference engine designed for deploying embodied AI models on real robots.
- Performance: end-to-end inference latency for π0.5 FP8 on Jetson Thor cut from 278ms to under 26ms (38.46Hz), enabling real-time robot control
- Stability: minimal Rust runtime with Python interface, memory-safe by design for long-running deployment
- Agility: built for Agentic Engineering — an agent system automatically explores and optimizes kernels, simplifying new model onboarding
First release supports π0.5 and WALL-OSS on Jetson Orin/Thor and RTX 4090. APXInf joins the RLinf ecosystem to bridge training and on-robot deployment; roadmap includes Qwen, Groot, nvfp4 optimization, and domestic chip support.
More from Embodied
- Wearing Whoop and Apple Watch on a single dual-slot band — tinyfool · 2026-11-02
- REA: async brain-cerebellum runtime for robots where policies never wait on the LLM — Pengfei Zhou · 2026-11-02
- Robot Combat Sports Coming to Japan with a Permanent Arena in Tokyo — cixliv · 2026-11-02
- REK robot fighting in the cage: humanoid bot in boxing gloves draws a crowd — broodsugar · 2026-11-02
- iPhone Duo: a folding iPhone mod impresses, but its hinge raises crackgate concerns — rudrank · 2026-11-02
- InflatableBots: University of Calgary's inflatable robots bring dynamic haptics to large-scale VR — lukas_m_ziegler · 2026-11-02