Honor robot “overtakes” a human for the first time: in Beijing, a humanoid ran a half marathon faster than the world record

A half-marathon in Beijing has become a showcase of how rapidly humanoid robotics is advancing. One of the participants — a humanoid robot — completed the 21 km distance in 50 minutes and 26 seconds. This result is faster than the current human world record of 57 minutes, set by Jacob Kiplimo.

Progress measured in hours and minutes

The year-on-year improvement is particularly striking. In 2025, the fastest robots needed around 2 hours and 40 minutes to complete the same distance. Within a year, engineers have reduced the completion time by more than threefold, reflecting major breakthroughs in motion control, balance, and energy efficiency of humanoid systems.

Two results, one manufacturer

The best official result was achieved by a robot developed by the Chinese company Honor. However, another robot from the same developer completed the course even faster — in 48 minutes and 19 seconds — but its result was not counted because it was remotely controlled.

The autonomous robot, finishing in 50:26, was declared the winner, as full self-operation without external control was the key requirement.

Autonomy as the main evaluation criterion

Organizers applied a scoring system in which the level of autonomy directly determines the final standing. As a result, the fully autonomous robot was awarded first place despite a faster, remotely operated counterpart.

An experimental competition format

According to the E-Town tech park in Beijing, around 40% of the robots operated in fully autonomous mode, while the remaining 60% were controlled remotely by human operators. This makes the event not only a sporting competition but also a testing ground for different levels of robotic autonomy.

Not all robots finished the race

Despite impressive results from the leaders, several humanoid robots failed to complete the course. Some lost pace, others stopped mid-race or exited stable movement patterns. This highlights that, despite rapid progress, autonomous running technologies are still in an active stage of development.