PyRUA-Lean Boosts GPT-6 Astra Robot Agents 14% Success with 65% Fewer Tokens
DAGroup-PKU · hf · 2026-10-02
PKU's DA Group introduces PyRUA-Lean, an interactive code-execution framework that slashes token overhead for VLM-based robot agents:
- Feedback-driven primitive composition: the agent composes classical robot primitives and learned VLA policies into Python cells with conditional checks and local retries, returning only explicitly requested images and state feedback for replanning.
- Results: across 700 simulated tasks from LIBERO-PRO, RoboTwin 2.0, and RoboCasa365, using the same GPT-6 Astra planner, overall success rises from 63.1% to 71.7% under equal LLM-call budgets; on instances both agents solve, it uses 49% fewer LLM calls and 65% fewer input tokens.
More from Embodied
- Removing human video drops robot OOD success from 60% to 7%: generalization comes from human video — gui_penedo · 2026-11-19
- PSYONIC's prosthetic Ability Hand now used by 90+ robotics orgs, raising Series A — DhowCIO · 2026-11-19
- Simulator misspecification breeds overconfidence: sim-to-real study on SPEED+ — Nikolay Dolgov · 2026-11-19
- Doppel world model renders the robot body and beats action-only baselines across 8 embodiments — mangahomanga · 2026-11-19
- FloMo: a world action model using 3D motion as the bridge from video pre-training to robot control — CSProfKGD · 2026-11-19
- MIT's paperclip-weight flying robot is first microrobot with onboard sensing and compute — MIT News AI · 2026-11-19