HumanoidPodium

2026-09-03

Ledger note: why 2025 and 2026 humanoid marks stay on separate rows

HumanoidPodium update: how we keep World Humanoid Robot Games and Beijing E-Town half-marathon clockings comparable without mixing rule years.

HumanoidPodium is a living ledger, not a one-shot dump. This note records how we treat the 2025 and 2026 public clockings already on the site, so readers and crawlers can see the table is still curated.

For the 100 m family, Tiangong Ultra's 2026 large-group stack is listed as separate rows: heat 9.39 s, semifinal 8.86 s (Reuters), and final 8.64 s (Associated Press). Those seconds live on the humanoid robot 100 m world record page and on the World Humanoid Robot Games 2026 event page. They are not World Athletics human records.

2025 Games 100 m times that used a published autonomy coefficient stay on their own rows. We do not rewrite them into the 2026 raw-final ranking. Empty cells still mean no sourced URL, not that a robot never raced.

Distance is the same rule with a different meet. The 2025 Beijing E-Town half marathon win (Tiangong Ultra, 2:40:42) and the 2026 autonomous half around 50 minutes (Honor Lightning and related rows) share a discipline family but not one ranking. Course, control mode, and year notes stay on the event and robot pages.

Football places from RoboCup Humanoid League 2021-2026 and the 2025 Games five-a-side final remain sourced places and scores only. Jump is still thin on purpose: one sourced standing high jump (Tiangong Ultra, 3.40 m) until more committee PDFs appear.

What changed this week on the product side: the public site now ships eight UI languages (English plus zh, ja, ko, de, fr, es, pt) with hreflang, and the open JSON/CSV snapshot on GitHub has matching README locales. The inclusion rule did not change: no public source_url, no row.

If you have a committee PDF or league results page we missed, use the contact page. We will not invent a time to fill a podium.

Related pages

Back to news