FULL STORY

X-Humanoid's Tiangong smashes sprint records at Beijing robot games

At Beijing's World Humanoid Robot Games, Tiangong won the 100m in 9.39s before setting 400m and 1500m marks far beyond human world records, overshadowing rivals like Honor's Lightning.

2026-08-22 ~ 2026-08-25 · 2 episodes · 24 posts

Episode 1 · Beijing Humanoid Robot Games 100m Dash Beats Bolt Record (2026-08-22, 17 posts)

At the World Humanoid Robot Games in Beijing, X-humanoid's 'Tiankung' crossed the finish line first in 9.39 seconds, while Honor's 'Lightning' recorded 9.32 seconds. Both times beat Usain Bolt's human world record of 9.58 seconds. The competition featured a 'fully autonomous' category where robots completed the race without human remote control. Videos of the event went viral on X, including a repost by Elon Musk, marking a widely circulated milestone for embodied AI.

Confirmed

  • Tiankung won with a time of 9.39 seconds; @CyberRobooo stated that Honor's Lightning recorded 9.32 seconds, noting its fast start but lack of explosiveness in the later stages
  • Both times beat Bolt's 9.58-second world record; @minchoi described Lightning's 9.32 seconds as a new speed record for Chinese robots
  • @breadli428 identified the event as the World Humanoid Robot Games 2026; @JFPuget, citing Reuters, confirmed the winning robot was developed by the Beijing Humanoid Robot Innovation Center
  • @csuwildcat reported that a Chinese humanoid robot achieved a 2.88-meter standing long jump at the same event, significantly surpassing Javier Sotomayor's human world record of 2.45 meters; performance improved about threefold from last year's best of 0.95 meters

Unconfirmed

  • There are discrepancies in reported times (9.3, 9.32, 9.39 seconds) across posts; the relationship between Lightning's 9.32 seconds and Tiankung's victory lacks clarification from primary sources, and official full results or test conditions are pending verification
  • @Polymarket referred to the champion robot as 'Tianzhuo', which conflicts with the majority of posts calling it 'Tiankung'
  • Details about the robot 'not falling apart at the waist and showing no sparks after hitting the safety wall' were only mentioned in retweets by users like @Scobleizer and lack original video evidence
  • @JFPuget mentioned discussions on social media regarding the authenticity and significance of the data, with speculation about manual engineering control systems, but these claims remain unverified

Why it matters

  • Multiple authors like @DrSingularity and @CyberRobooo view this as the first widely publicized case of robots 'surpassing human limits' in motion control and physicality; @csuwildcat highlighted the rapid progress, noting the triple jump in long jump performance within one year
  • The completion of the race without human remote control, along with structural stability and steering capabilities at high speeds, demonstrates actual advancements in hardware and motion control algorithms

Episode 2 · Tiangong Humanoid Robot Smashes Human 400m and 1500m World Records (2026-08-23, 7 posts)

At the second World Humanoid Robot Games (WHRG'26) held in Beijing, a humanoid robot named "Tiangong" ran 400 meters in 38.15 seconds and 1,500 meters in 2 minutes 21.6 seconds, dramatically breaking world records held by humans—the 400m record was 43.03 seconds by South African athlete Wayde van Niekerk, and the 1500m record was 3 minutes 26 seconds.

Confirmed

  • The 400m and 1500m results and the corresponding prior human records appeared consistently across multiple posts
  • The event was the second World Humanoid Robot Games, held in Beijing

Why it matters

  • Slow-motion footage of the 400m champion spread widely on social platforms: @herbiebradley reposted the slow-motion clip; @RealGeneKim relayed that its momentum-based propulsion is almost as groundbreaking as AlphaGo's Move 37, with some even arguing humans cannot run this way
  • @basedjensen reposted that its running form resembles the classic anime trope of sprinting to school with toast in one's mouth, sparking a lighthearted discussion on bipedal robot gait optimization
  • The result demonstrates significant progress in high-speed bipedal locomotion and gait control for humanoid robots