Humanoid robots set four robot-specific athletics records at the 2026 World Humanoid Robot Games in Beijing, highlighting rapid advances in robotic speed, balance, coordination and mechanical design. Tiangong Ultra, a humanoid robot developed by the Beijing Humanoid Robot Innovation Center, won the 100-meter, 400-meter and 1,500-meter races, while an X-Humanoid robot recorded the event’s highest standing high jump.
Although these performances are not official replacements for human athletics records, they demonstrate how quickly humanoid robots are improving under specialized competition rules and conditions. The five-day event ran from August 22 to 26 at Beijing’s National Speed Skating Oval and featured 2,056 robots from 666 teams representing 16 countries.
The 2026 competition included track and field events, soccer, martial arts, warehouse logistics, factory-style assembly, object handling and autonomous charging tasks. More than 40% of the events required robots to operate without direct human control.
Humanoid Robots Set Four Records at the 2026 Robot Games
The robot athletics events were designed specifically for machines, meaning their results must be compared carefully with human records. Even so, the performances reveal major improvements in robotic actuators, balance control, software and energy management.
100 meters: Tiangong Ultra — 8.64 seconds
Human benchmark: Usain Bolt’s official 100-meter world record is 9.58 seconds, set at the 2009 World Athletics Championships in Berlin.
Robot result: Tiangong Ultra won the men’s 100-meter final in 8.64 seconds. The robot had already run 9.39 seconds in an earlier race, improved to 8.86 seconds during qualifying and then lowered its time again in the final.
Tiangong Ultra competes in a robot athletics event.
(Image credit: X-Humanoid)
Tiangong Ultra Wins the 400-Meter Race
400 meters: Tiangong Ultra — 38.15 seconds
Human benchmark: Wayde van Niekerk holds the official human 400-meter world record of 43.03 seconds, set at the 2016 Rio Olympics.
Robot result: Tiangong Ultra also won the 400-meter final in 38.15 seconds. The robot’s time was nearly five seconds faster than the human benchmark, although the event used robot-specific rules and conditions that make direct comparisons unsuitable.
Humanoid robots compete in the 4×100-meter relay final at the 2026 World Humanoid Robot Games in Beijing.
(Image credit: VCG via Getty Images)
Robot Breaks the 1,500-Meter Competition Record
1,500 meters: Tiangong Ultra — 2 minutes 21.6 seconds
Human benchmark: Morocco’s Hicham El Guerrouj holds the official men’s 1,500-meter world record of 3 minutes 26 seconds, set in Rome in 1998.
Robot result: Tiangong Ultra won the robot 1,500-meter final in 2 minutes 21.6 seconds. The next two competitors also finished ahead of the human benchmark, recording times of 2:30.00 and 2:30.22 under the event’s robot-specific conditions.
Tiangong Ultra runs in the 1,500-meter final at the World Humanoid Robot Games in Beijing.
(Image credit: Kevin Frayer/Stringer via Getty Images)
X-Humanoid Robot Records a 2.88-Meter Standing High Jump
Standing high jump: 2.88 meters, or 9.45 feet
Human benchmark: Javier Sotomayor of Cuba holds the official human high-jump record at 2.45 meters, set in Salamanca, Spain, in 1993.
Robot result: An X-Humanoid robot cleared 2.88 meters in a standing high jump, surpassing the previous robot record of 0.95 meters established at the first competition in 2025.
However, the robot event was not identical to the human high jump. Human athletes use a running approach, a single-leg takeoff and the Fosbury flop to convert horizontal speed into vertical lift. The robot competition involved a standing jump, which relies on a different combination of motors, actuators and control software.
What the Humanoid Robot Records Reveal
The results from Beijing show that humanoid robots are becoming more capable in highly specialized physical tasks. Sprinting and jumping require powerful actuators, precise foot placement, rapid balance corrections and software that can respond to small errors in real time.
Scott Walter, director of robotics research at RoboStrategies, said the record performances were largely the result of mechanical optimization for individual events. Designing a machine to perform one task extremely well can produce impressive results, even if the robot has limited abilities outside that task.
Reducing unnecessary components can lower a robot’s weight and decrease the number of potential failure points. In some competition machines, degrees of freedom in the arms, hips and ankles were limited to improve movement efficiency and allow faster motion.
That specialization also creates important limitations. A robot built for straight-line acceleration may struggle to slow down, navigate curves, stop safely or recover from unexpected obstacles. Some competitors fell, collided with barriers or experienced mechanical failures during the games.
Robot teams also faced challenges in relay events. Machines without functional hands may be unable to complete a traditional baton handoff, showing why success in one athletic event does not necessarily translate into broader humanoid abilities.
Why General-Purpose Humanoid Robots Remain Difficult
The 2026 World Humanoid Robot Games also tested practical skills, including connecting charging cables, moving objects, using tools and completing simulated industrial tasks. These challenges require more than speed or strength.
A capable general-purpose robot must identify an object, estimate its position, select the correct grip, apply the right amount of force and recover when conditions change. Reliable manipulation remains one of the biggest challenges because robots still lack the dexterity, strength and sensitivity of human hands.
As Walter noted, specialization is relatively straightforward compared with creating a robot that can perform many different tasks reliably. A future decathlon-style event could provide a tougher test by combining speed, endurance, power, coordination, robustness and dexterity.
The 2026 games therefore offer two important lessons: humanoid robots can now achieve remarkable results in carefully optimized events, but dependable performance across unpredictable real-world tasks remains the greater challenge.
Source: www.livescience.com