Physical AI

Show-Harness: Just a VLM Agent Can Play Robots

Foundation vision-language models (VLMs) exhibit broad intelligence about the world, yet translating this intelligence into robot control remains challenging. We present Show-Harness, an Embodied Harness that enables VLMs to "play" robots through a compact semantic interface linking intent to action. Show-Harness exposes discrete semantic action units that VLMs can naturally reason over, while embodiment-specific interpreters deterministically ground them into local robot actions, keeping the VLM directly responsible for fine-grained physical decisions. Through the same interface, Show-Harness demonstrates the feasibility of (1) directly unlocking closed-source frontier VLMs for zero-shot robot control, and (2) adapting small-scale open-source VLMs for low-cost deployment with just a few GPU-hours of fine-tuning. We further develop GUMI (GUI Manipulation Interface), which extends the same semantic action space to GUI-based demonstration collection, allowing humans and agents to "play" robots across embodiments without specialized teleoperation hardware. Extensive experiments show that Show-Harness-equipped VLM agents generalize robustly across tasks, embodiments, and environments, outperforming representative agentic and VLA paradigms. These results suggest that the right interface can unlock substantial embodied capability from foundation VLMs, without requiring additional model capacity or costly embodiment-specific pretraining.

Yanzhe Chen, Zechen Bai, Zhijun Cao

Fire3D Turns RGB Video into Simulation-Ready 3D Scenes

Fire3D converts an unsegmented image or casual RGB video into a complete textured, simulation-ready 3D scene in under a minute. It discovers objects and estimates their 6-DoF poses without supplied masks, boxes or prompts.

StereoPolicy Learns 3D Structure Directly From Stereo Images

StereoPolicy removes the need for depth sensors, depth maps, point clouds, and LiDAR by learning 3D structure directly from stereo image pairs. Across five real-world tabletop manipulation tasks, it reaches 59% success compared with 42% for RGB, 41% for RGB-D, and 14% for PointNet.

Bimo: An Open-Source Biped You Can Run From a PC, an SBC or Its Own RP2040

Bimo, an open-source bipedal robot, can be driven from a PC, fitted with an SBC for onboard compute, or run its walking policy directly on the RP2040. The repo covers API, IsaacLab training, MCU firmware and ROS2.

GPT-6 Astra Makes a Unitree Go1 Climb Stairs Blind

GPT-6 Astra drives a Unitree Go1 up stairs in simulation: instead of writing gait code it emits joint-angle sequences directly, so this is blind locomotion with physics paused between inference calls. The author ties it back to the 2023 GPT-4 trajectory-generation work.

GPT-6 Astra Drives a Robot Arm From a Human Video on the First Pass

Drop a recording of a human doing a novel task into the codex app, prompt it to drive a robot arm the same way, and GPT-6 Astra works on the first pass: physical in-context learning out of the box.

SkillX: one RL policy that dribbles, traps and shoots on a real Noetix E1

Humanoid soccer makes one robot balance, move, kick a ball and switch skills over long horizons, and most prior sports pipelines still train each behaviour as a separate stage. SkillX (Noetix Robotics + Tsinghua University, CoRL 2026) instead learns dribbling, trapping and shooting inside a single command-conditioned RL policy. Three designs hold the shared actor together: skill-specific adversarial motion priors give per-skill style rewards, skill-specific critics estimate skill-dependent values, and an object-aware temporal encoder compresses history including ball state. On a real Noetix E1 it chains dribble-into-shoot over long rollouts, deployed first with motion capture supplying base and ball pose, then with onboard vision only (head-mounted ZED2i, visual-inertial odometry, YOLOv8 ball detection). Remove object observations and rewards and the same framework still produces diverse whole-body skills and transitions.

RoboSim: A Robot-Native Rust Simulation Engine

A Rust robot-native engine: deterministic headless simulation core with wgpu rendering, Rapier physics behind backend-neutral traits plus a MuJoCo adapter, LiDAR/RGB-D/IMU sensors, Python bindings and an optional ROS 2 bridge. Showcases cover Unitree Go2 quadruped and G1 23-DoF humanoid locomotion, mobile manipulation in a captured 3DGS interior, vehicle dynamics at the grip limit and PLATEAU city traffic. The clip is a headless CLI run: 600 steps at 60 Hz with a determinism check and a replay digest.

How to Build a Humanoid Robot: A Seven-Layer Constraint Chain from Joint Modules to Sim-to-Real

A humanoid robot is not an assembly of seven modules but a stack of physics equations that set each other's boundary conditions. This article computes the whole-machine stack layer by layer: joint motor modules (declared torque versus real quasi-static CoP demand — knee margins across three vendors converge to 2.23-2.42x while BHL's knee has only 1.51x) -> IMU (lever-arm pseudo-acceleration is 21,752x the sensor noise floor, so mounting position matters four orders of magnitude more than the datasheet) -> materials and structure (three BOM revisions of AgiBot's X1 as a load-path history: every part entering the closed-chain drivetrain upgraded to 7075-T6 / TC4 / 17-4PH) -> sensors (fix the observation space before the shopping list) -> battery and BMS ('all joints at peak simultaneously' is physically impossible: G1's 46,062 W against a 421 Wh pack is 109 C) -> software control and the CAN-level low-side boards (22 nodes at 500 Hz on one bus is 130.9% load, so it must be split into four) -> simulation training and sim-to-real: domain randomization, sim2sim, zero calibration (ATOM01's 2.093 rad waist-yaw assembly offset, the |q| < 1e-2 rad acceptance gate, and write_motor_flash() being a no-op in three of the four motor drivers), plus 10 of 13 real failure modes being hardware calibration rather than simulation fidelity. Every figure comes from programmatic parsing of the five machines' public model files, deployment and calibration source, plus official vendor specifications, and is recomputable.

TANGO: Humanoid Navigation in Cluttered Environments with a Whole-Body Vision-Language-Action Model

We study the problem of navigating cluttered indoor environments with a humanoid robot. Unlike conventional methods that model navigation as a 2D path planning problem, humanoid traversal in cluttered environments requires continuous geometry-aware whole-body adaptation, including coordinated arm placement, torso adjustment, and gait modulation for collision-free movement through complex 3D spaces. We introduce TANGO, the first whole-body vision-language navigation framework for language-conditioned humanoid traversal in cluttered environments. Given a natural-language instruction and egocentric RGB observations, TANGO directly predicts 29-DoF joint-space actions for downstream whole-body control. We train TANGO entirely in simulation by synthesizing diverse collision-free traversal behaviors via global path planning, kinematic whole-body motion generation, obstacle-aware motion editing, and RL-based tracking. This pipeline provides dynamically feasible action supervision for learning language-conditioned whole-body policies. In extensive simulation experiments, TANGO demonstrates state-of-the-art performance in vision-language navigation, while outperforming strong modular baselines in navigating challenging scenes requiring obstacle negotiation. Lastly, we deploy TANGO zero-shot on a Unitree G1 humanoid robot, and observe robust language-guided traversal in cluttered real-world scenes without training on any real-world navigation data.

Anqi Li, Yuxin Chen, Zhaobo Li

PGMT: Perceptive General Motion Tracking for Humanoid Robots

Humanoid motion trackers can reproduce diverse whole-body motions, but their performance degrades on complex terrain where terrain-agnostic references become physically infeasible. We present PGMT, a Perceptive General Motion Tracking pipeline for humanoid robots that learns terrain adaptation from independently selected motion references and terrains. PGMT first learns a general tracking and recovery prior, then incorporates terrain perception through motion-conditioned terrain glimpses that selectively encode regions relevant to the current motion. Terrain-aware tracking relaxation allows necessary deviations from the reference while preserving its motion intent. Zero-shot deployment on a Unitree G1 demonstrates robust terrain-adaptive locomotion and whole-body motion execution over real-world terrain with obstacles up to 37 cm high, while supporting teleoperation, dynamic motion tracking, and fall recovery. PGMT extends general humanoid motion tracking beyond flat ground, providing a unified policy for terrain-adaptive locomotion, diverse whole-body behaviors, and teleoperation in complex environments.

Hongyi Li, Li Peizhuo, Yucheng Tao

GPT-6 Runs Real2Sim End to End from Robot Demonstrations

Given only multi-view RGB, robot actions and one prompt, GPT-6 calibrated the cameras, built the object assets, ran physics system identification, executed the scene in MuJoCo and rendered it in Blender: an end-to-end real2sim pipeline rebuilt from robot demonstrations.