Humanoid Robot Leaderboard
156 robots · ranked byXHI
| # | Robot | Status | Autonomy | XHI |
|---|---|---|---|---|
| 1 | 🇺🇸 Figure 03Figure AI |
commercial | task-autonomous | 87.0 |
| 2 | 🇺🇸 Atlas (electric)Boston Dynamics |
pilot | task-autonomous | 84.8 |
| 3 | 🇩🇪 Agile ONEAgile Robots |
pilot | task-autonomous | 81.2 |
| 4 | 🇨🇳 L7RobotEra (Beijing Xingdong Jiyuan / 星动纪元) |
commercial | task-autonomous | 80.4 |
| 5 | 🇺🇸 NEO (Gamma)1X Technologies |
limited-production | supervised-autonomy | 80.4 |
| 6 | 🇨🇳 Iron (next-gen)XPeng |
pilot | supervised-autonomy | 80.2 |
| 7 | 🇺🇸 Figure 02Figure AI |
retired | task-autonomous | 76.6 |
| 8 | 🇨🇳 Walker S2UBTech Robotics |
limited-production | task-autonomous | 75.5 |
| 9 | 🇨🇳 H2 PlusUnitree Robotics |
commercial | task-autonomous | 74.5 |
| 10 | 🇨🇳 AlphaBot 2AI2 Robotics |
commercial | task-autonomous | 74.4 |
| 11 | 🇺🇸 ApolloApptronik |
pilot | supervised-autonomy | 73.8 |
| 12 | 🇩🇪 4NE-1NEURA Robotics |
pilot | supervised-autonomy | 72.7 |
| 13 | 🇺🇸 THEMIS V2Westwood Robotics |
commercial | task-autonomous | 72.5 |
| 14 | 🇺🇸 Optimus Gen 2Tesla |
pilot | supervised-autonomy | 72.2 |
| 15 | 🇨🇳 Kuavo-5Leju Robot |
commercial | task-autonomous | 72.1 |
| 16 | 🇺🇸 Optimus Gen 3Tesla |
pilot | supervised-autonomy | 70.1 |
| 17 | 🇨🇳 G1Unitree Robotics |
commercial | task-autonomous | 69.8 |
| 18 | 🇨🇳 Galbot G1Galbot |
commercial | task-autonomous | 69.7 |
| 19 | 🇨🇭 AEONHexagon Robotics |
pilot | supervised-autonomy | 69.0 |
| 20 | 🇺🇸 Figure 01Figure AI |
retired | task-autonomous | 69.0 |
| 21 | 🇨🇳 Tiangong 2.0 PlusX-Humanoid (Beijing Humanoid Robot Innovation Center) |
limited-production | task-autonomous | 67.9 |
| 22 | 🇨🇦 Phoenix (Gen 7)Sanctuary AI |
pilot | supervised-autonomy | 67.3 |
| 23 | 🇺🇸 Optimus (Gen 2/3)Tesla |
pilot | task-autonomous | 66.9 |
| 24 | 🇨🇳 Forerunner K2Kepler Robotics |
commercial | supervised-autonomy | 66.8 |
| 25 | 🇫🇷 Calvin-40Wandercraft |
commercial | task-autonomous | 66.0 |
Reading the table. XHI is the weighted composite (0–100)
for this category, normalised against humanoid-specific reference scales.
“Autonomy” shows the demonstrated mode — teleoperated
systems are capped low regardless of hardware. “Conf.” is the data-confidence grade for that
robot's specs. Click any row for the full breakdown and sources.
These rankings are not comparable across categories. Each category has
its own index, its own pillar weights and its own reference scales, so an XHI of 71
and a score of 71 on another leaderboard answer different questions. Only autonomy and commercial
readiness are scored from ladders that mean the same thing everywhere.
Why →