RPent open-sourced: 92.6% on LIBERO-PRO and 7x faster embodied agent execution
量子位 · wechat · 2026-09-21
RPent, an embodied-agent infrastructure open-sourced by a team from Tsinghua, Wuxin Wuqiong and Zhengxing Innovation, connects LLM planning with VLA expert models, memory and robot interfaces into a closed loop for real-world tasks.
- Modular architecture: LLM handles planning, VLA/WAM models handle fine manipulation, a three-layer memory system deposits validated experience (disturbance-task success rose from 31.0% to 87.0%)
- Unified MCP/RPC/MHS interfaces decouple decisions from execution; supports LIBERO/RoboCasa sims and Franka, YAM robots
- FlashMode with TaskCards compresses verified task flows into reusable cards: on LIBERO Object, execution time dropped from 283.6s to 40.9s (7x) with only a 3.5-point success drop
- RPent/GPT-6 Astra scores 92.63% on LIBERO-PRO, 10.23 points above second place; based on the team's July HarnessVLA work
Code: github.com/RLinf/RPent
More from Embodied
- Asking for the best way to run MiniMax H3 video generation on a DGX Spark — Shady-Dragon · 2026-09-21
- Meta reportedly building camera-free smart glasses codenamed Luna for privacy — emmanuelvivier · 2026-09-21
- AgiBot tops global humanoid shipments in H1; Chinese firms hold over 97% of market — emmanuelvivier · 2026-09-21
- Chinese T800 humanoid patrols Shenzhen streets with SWAT, priced from $25,000 — TinfoilTricorn · 2026-09-21
- Agent designs its own hardware: Astra produces USB-powered 4-light PCB files — paraschopra · 2026-09-21
- Starlink enables West Africa's first telesurgical kidney removal across 500 km — elonmusk · 2026-09-21