Robots Beat Human Half-Marathon Record

In April 2026, more than 100 humanoid robots competed in a half-marathon event held in an innovation zone on the outskirts of Beijing. A robot from Chinese smartphone maker Honor won the race with a time of 50 minutes and 26 seconds — a result that not only outpaced all other robotic competitors but also surpassed the human half-marathon world record of 57 minutes and 31 seconds.

The event marks a significant milestone in the athletic capabilities of humanoid robots. Just a few years ago, robots still struggled to navigate uneven terrain without stumbling. Today, they can complete a 21-kilometre course at a steady, competitive pace.

100m Sprint Record Within Reach

Even more remarkably, humanoid robots are making rapid progress in the 100-metre sprint. Chinese robotics company Unitree Robotics released video showing its H1 humanoid robot reaching a peak sprint speed of 10 metres per second — closing in on Usain Bolt’s human world record of 9.58 seconds (approximately 10.4 m/s).

New Scientist reports that robots are not only outperforming humans in endurance events but are rapidly narrowing the gap in speed events as well. Analysts suggest that, at the current rate of development, humanoid robots could break the human 100m world record within the next few years.

Technology Drivers

This progress is driven by breakthroughs across multiple technology domains. Advanced motor drivers, lightweight carbon-fibre materials, and AI-powered motion control algorithms enable modern robots to achieve far more agile and efficient movement.

Organisers of the Beijing half-marathon noted that competing robots must complete the race in fully autonomous mode, without any external control. This requires real-time environmental perception, obstacle avoidance, and dynamic balance adjustment — capabilities that were unimaginable in robotics just a few years ago.

Industry Significance

The leap in humanoid robot athletic ability is more than a technological showcase; it carries profound industrial implications. Robots capable of running stably on complex terrain are better equipped for real-world applications in disaster relief, exploration, logistics, and more.

However, experts caution that performance on a standardised track still differs from real-world deployment. Maintaining stability and adaptability in unstructured environments remains a core challenge that the robotics industry must address before widespread adoption becomes feasible.

Source: New Scientist, NPR