The Unitree G1 has quickly become one of the most significant milestones in the 2024–2026 robotics era. Standing at roughly 1.32m (4’4″) and weighing about 35kg, it is designed to be a high-performance, mass-produced humanoid that bridges the gap between lab research and real-world application.
Core Technical Specifications for all models
While the internals vary, the physical frame of the G1 remains consistent across all versions:
| Feature | Specification |
| Height | 1320mm (Standing) / 690mm (Folded) |
| Weight | ~35kg (includes 9000mAh Quick-release battery) |
| Battery Life | ~2 hours (Active) / 3-4 hours (Standby) |
| Max Speed | 2.0 m/s (approx. 7.2 km/h) |
| Perception | 3D LiDAR (Livox Mid-360) + Depth Camera (Intel RealSense D435i) |
| Audio | 4-Microphone Array + 5W Stereo Speakers |
| Connectivity | Wi-Fi 6, Bluetooth 5.2 |
Version Comparison: Basic vs. EDU
The most critical distinction is between the G1 Basic and the G1 EDU series.
Warning: The G1 Basic cannot be upgraded to the EDU version later due to fundamental differences in motor mounts, internal wiring, and compute slots.
1. Unitree G1 Basic
The “Entry” model, often referred to as a Demo Bot. It is designed for showcases and basic interaction rather than development.
Degrees of Freedom (DoF): 23 fixed.
Knee Torque: 90 N·m.
Arm Payload: ~2kg.
Compute: 8-core High-performance CPU only.
Software: Closed system (No SDK access; controlled via App/Remote).
2. Unitree G1 EDU Series
The EDU versions are the “Pro” grade, featuring 120 N·m knee torque, a 3kg arm payload, and an NVIDIA Jetson Orin module (up to 275 TOPS) for on-board AI processing. They offer Full SDK Access (Python, C++, ROS2).
| Version | Total DoF | Hand Type | Key Differentiator |
| EDU Standard (U1) | 23 | Simple Gripper | Base level for ROS2 research. |
| EDU Plus (U2) | 29 | Simple Gripper | Adds 3-axis waist and 7-DoF arms. |
| EDU Ultimate (U3/U4) | 43 | Dex3-1 Hands | Includes 3-finger force-controlled hands. |
| EDU Ultimate D (U7) | 43 | Tactile Hands | Adds 33 tactile sensors per hand for delicate tasks. |
| EDU Anthropomorphic | 41+ | 5-Finger Hands | Uses BrainCo or Inspire 5-finger hands for social robotics. |
Applications & Use Cases
🔬 Research & Development
Locomotion Mastery: Researchers use the G1 to develop reinforcement learning models for walking on uneven terrain, stair climbing, and recovery from falls.
Manipulation Studies: The EDU Ultimate versions allow labs to study “Hand-Eye Coordination” using the Dex3-1 hands to pick up tools, open bottles, or solder electronics.
🏠 Domestic & Service Tasks
Home Assistance: As seen in recent 2025/26 demos, the G1 is being trained for “tidying up,” watering plants, and basic surface cleaning.
Retail/Hospitality: Serving as a high-tech “Greeter” or guide in malls and corporate offices.
⚠️ Hazardous Environments
Inspection: Replacing humans in zones with chemical leaks or structural instability. Its 3D LiDAR allows it to map dark or smoke-filled environments autonomously.
Remote Handling: Using a VR interface, operators can “teleoperate” the robot to handle dangerous materials using the dexterous hands.
🎓 Education & STEM
University Labs: Teaching the next generation of engineers how to program humanoid kinematics and computer vision using the ROS2 ecosystem.
Interested in learning more about Unitree G1 Humanoids?
Check all Unitree Humanoid versions in our store.
