// Research · Scorekeeper, not cheerleader
Humanoid Research
What the science actually shows about humanoid robots, in plain English. Each week we read the new papers, check the numbers against the source, and say what was tested on a real robot and what was only simulated.
Updated November 13, 2026
Where things stand
Can humanoid robots walk reliably yet?
On set lab courses, increasingly yes.
3 papers · Where things stand →Hands and handling objectsHow well can humanoid robots pick up and handle objects?
Grasping is getting good in controlled tests; fine finger work is still early.
3 papers · Where things stand →NavigationCan humanoid robots find their way around on their own?
Within a room, following directions in plain language, yes, in demonstrations.
1 paper · Where things stand →Working with peopleAre humanoid robots safe to work alongside people?
Researchers are building the pieces, but the evidence is lab-scale.
3 papers · Where things stand →New this week
What it means for the Robot Rankings
Capability is 25% of each company's score in the weekly Robot Rankings, and we check capability claims against published research like this. None of the papers here came from a ranked company. Unitree's G1 was the test robot in 5 of the 10 papers here, a sign of how widely it is used as a research platform.
How we pick and rate papers
- We cover primary research that tests AI, perception, planning or control on a humanoid robot, or reports a distinct humanoid experiment. Product launches, funding news and promotional videos without a paper are left out.
- Every paper links to its primary source. Most are arXiv preprints, which are not peer-reviewed; we say so, and add the journal or conference version when one is confirmed.
- We check the numbers against the paper itself, and say whether we read the abstract or the paper text.
- We separate physical-robot results from simulation, and note when a robot was teleoperated rather than acting on its own.
- Strong means quantified tests on a physical humanoid with defined tasks and real comparisons. Moderate means limited hardware results or partial evidence. Preliminary means simulation-only or exploratory work.
Spotted an error, or a paper we should cover? Tell us, with a link to the paper.