A humanoid robot of easily assembled 3-D-printed parts and off-the-shelf components costs less than $5,000 and just might be the model for would-be robotics beginners to jump into the field.
Berkeley Humanoid Lite, the low-cost, fully open-source humanoid robot designed by engineers at the University of California at Berkeley, is 3.3 feet tall and weighs about 35 pounds, can be built in about a week by researchers and developers seeking a practical and customizable platform. Its 3-D-printed actuators, built with cycloidal gears to increase durability, have held up under heavy use, including walking and object manipulation tasks.
A modular 3-D-printed gearbox for actuators and robot body is the core piece of the design. Large gear teeth reduce wear and increase longevity by distributing load across a larger surface area. Remaining components can be sourced from widely available e-commerce platforms or fabricated with standard desktop 3-D printers, keeping costs down. Replacements for broken or worn parts can be easily fabricated using a 3-D printer.
The project, presented at the 2025 Robotics Science and Systems conference in Los Angeles in June, offers some dazzling capabilities. The Berkeley demonstrated two control systems: joystick-based teleoperation for tasks such as handling a Rubik’s Cube, and reinforcement learning-based walking, trained in simulation and transferred to the robot using zero-shot policy transfer.
“I believe in the spirit of open-source communities, an ecosystem where people share ideas and knowledge. We hope that Berkeley Humanoid Lite will help move us closer to democratizing the development of humanoid robotics,” Yufeng Chi, a Ph.D. student in the Department of Electrical Engineering and Computer Sciences who is part of the team of Berkeley engineers, said in a statement.
The affordable humanoid robotics project, one of several nationwide alongside efforts at Carnegie Mellon University and elsewhere, could have far-reaching implications as multiple industries map their physical artificial intelligence (AI) strategies.
Humanoid robotics have kicked around for years, slowly evolving as devices for use in industrial automation, health care, research, and personal assistance.
But their applications at work and in homes has been stultified by complexity and cost, which has limited commercially available hardware, say Chi and others. In the interim, simpler robots like drones and wheeled machines dominate the field because they’re easier and cheaper to build.
Engineers at Berkeley hope their humanoid quickly accelerates the field. They’ve built an online community where others can collaborate, troubleshoot, and improve on their design.
“Our aim is to help people in the research and education space become familiar with how a humanoid system works, including how to assemble and develop a robot platform, by providing them with a template to get started,” Chi said. “It’s all about looking into an example system and building this system from scratch, one block after another. Then after you’ve gained experience and confidence, you can build upon it and take it to a new level.”
