Public evidence coverage
Humanoid robot disclosure index
How much of the catalog has a sourced specification, price, official buy path or document link? Missing entries mean not found in the public sources we checked.
378 robot records · 42 eligible makers · Data last verified 2026-10-13
Coverage by specification
All catalog models, including historical and torso-only robots, share the same denominator.
- Height291 / 378 robots
- Weight275 / 378 robots
- Total DoF276 / 378 robots
- DoF per hand141 / 378 robots
- Payload169 / 378 robots
- Maximum speed148 / 378 robots
- Battery107 / 378 robots
- Runtime188 / 378 robots
- Charge time50 / 378 robots
- Peak joint torque74 / 378 robots
- Swappable battery64 / 378 robots
- Actuation153 / 378 robots
- Hand / end effector218 / 378 robots
- Onboard compute177 / 378 robots
- Sensors266 / 378 robots
- AI stack130 / 378 robots
Maker coverage ranking
At least three catalog robots per maker. Each model has 19 equally weighted coverage slots.
| Rank | Maker | Robots | Spec values / possible | Prices / robots | Buy paths / robots | Documents / robots | Found / possible | Score |
|---|---|---|---|---|---|---|---|---|
| 1 | EngineAI Robotics | 4 | 48 / 64 | 4 / 4 | 4 / 4 | 1 / 4 | 57 / 76 | 75.0% |
| 2 | Noetix Robotics | 3 | 37 / 48 | 2 / 3 | 2 / 3 | 0 / 3 | 41 / 57 | 71.9% |
| 3 | AgiBot (Zhiyuan Robotics) | 7 | 82 / 112 | 2 / 7 | 6 / 7 | 1 / 7 | 91 / 133 | 68.4% |
| 4 | Booster Robotics | 3 | 34 / 48 | 1 / 3 | 3 / 3 | 1 / 3 | 39 / 57 | 68.4% |
| 5 | ROBOTIS Co., Ltd. | 5 | 56 / 80 | 3 / 5 | 3 / 5 | 2 / 5 | 64 / 95 | 67.4% |
| 6 | Unitree Robotics | 9 | 97 / 144 | 4 / 9 | 8 / 9 | 4 / 9 | 113 / 171 | 66.1% |
| 7 | PAL Robotics | 3 | 32 / 48 | 0 / 3 | 2 / 3 | 3 / 3 | 37 / 57 | 64.9% |
| 8 | Fourier | 5 | 56 / 80 | 1 / 5 | 0 / 5 | 2 / 5 | 59 / 95 | 62.1% |
| 9 | CASBOT | 3 | 34 / 48 | 0 / 3 | 1 / 3 | 0 / 3 | 35 / 57 | 61.4% |
| 10 | Galbot | 3 | 30 / 48 | 1 / 3 | 3 / 3 | 0 / 3 | 34 / 57 | 59.6% |
| 11 | UBTech Robotics | 13 | 130 / 208 | 3 / 13 | 11 / 13 | 0 / 13 | 144 / 247 | 58.3% |
| 12 | LimX Dynamics | 4 | 37 / 64 | 3 / 4 | 3 / 4 | 1 / 4 | 44 / 76 | 57.9% |
| 13 | 1X Technologies | 3 | 31 / 48 | 1 / 3 | 1 / 3 | 0 / 3 | 33 / 57 | 57.9% |
| 14 | KAIST (Korea Advanced Institute of Science and Technology) | 3 | 30 / 48 | 1 / 3 | 0 / 3 | 0 / 3 | 31 / 57 | 54.4% |
| 15 | MagicLab | 4 | 38 / 64 | 0 / 4 | 3 / 4 | 0 / 4 | 41 / 76 | 53.9% |
| 16 | Beijing Humanoid Robot Innovation Center (X-Humanoid) | 6 | 53 / 96 | 2 / 6 | 4 / 6 | 1 / 6 | 60 / 114 | 52.6% |
| 17 | Kepler Robotics | 4 | 33 / 64 | 2 / 4 | 1 / 4 | 0 / 4 | 36 / 76 | 47.4% |
| 18 | RobotEra | 4 | 33 / 64 | 1 / 4 | 1 / 4 | 1 / 4 | 36 / 76 | 47.4% |
| 19 | Figure AI | 3 | 27 / 48 | 0 / 3 | 0 / 3 | 0 / 3 | 27 / 57 | 47.4% |
| 20 | National Institute of Advanced Industrial Science and Technology (AIST) | 6 | 51 / 96 | 1 / 6 | 0 / 6 | 0 / 6 | 52 / 114 | 45.6% |
| 21 | Karlsruhe Institute of Technology (KIT) | 3 | 26 / 48 | 0 / 3 | 0 / 3 | 0 / 3 | 26 / 57 | 45.6% |
| 22 | Galaxea AI | 3 | 21 / 48 | 1 / 3 | 3 / 3 | 0 / 3 | 25 / 57 | 43.9% |
| 23 | Leju Robotics | 7 | 57 / 112 | 0 / 7 | 0 / 7 | 1 / 7 | 58 / 133 | 43.6% |
| 24 | Digit Robotics | 3 | 21 / 48 | 0 / 3 | 3 / 3 | 0 / 3 | 24 / 57 | 42.1% |
| 25 | Boston Dynamics | 3 | 21 / 48 | 0 / 3 | 1 / 3 | 1 / 3 | 23 / 57 | 40.4% |
| 26 | Zhejiang Humanoid Robot Innovation Center | 3 | 18 / 48 | 0 / 3 | 3 / 3 | 0 / 3 | 21 / 57 | 36.8% |
| 27 | Waseda University | 7 | 45 / 112 | 0 / 7 | 0 / 7 | 0 / 7 | 45 / 133 | 33.8% |
| 28 | University of Tehran | 5 | 31 / 80 | 1 / 5 | 0 / 5 | 0 / 5 | 32 / 95 | 33.7% |
| 29 | Honda Motor Co., Ltd. | 6 | 37 / 96 | 0 / 6 | 0 / 6 | 0 / 6 | 37 / 114 | 32.5% |
| 30 | AKINROBOTICS | 6 | 34 / 96 | 1 / 6 | 1 / 6 | 0 / 6 | 36 / 114 | 31.6% |
| 31 | EIR Technology | 3 | 16 / 48 | 2 / 3 | 0 / 3 | 0 / 3 | 18 / 57 | 31.6% |
| 32 | Shanghai Humanoid Robot Innovation Center | 4 | 22 / 64 | 0 / 4 | 0 / 4 | 0 / 4 | 22 / 76 | 28.9% |
| 33 | AI2 Robotics | 3 | 15 / 48 | 0 / 3 | 1 / 3 | 0 / 3 | 16 / 57 | 28.1% |
| 34 | Tesla, Inc. | 3 | 16 / 48 | 0 / 3 | 0 / 3 | 0 / 3 | 16 / 57 | 28.1% |
| 35 | Hiroshi Ishiguro Laboratory | 4 | 20 / 64 | 0 / 4 | 0 / 4 | 0 / 4 | 20 / 76 | 26.3% |
| 36 | KEENON Robotics | 3 | 15 / 48 | 0 / 3 | 0 / 3 | 0 / 3 | 15 / 57 | 26.3% |
| 37 | University of Tokyo | 4 | 19 / 64 | 0 / 4 | 0 / 4 | 0 / 4 | 19 / 76 | 25.0% |
| 38 | Shanghai Electric Group | 3 | 14 / 48 | 0 / 3 | 0 / 3 | 0 / 3 | 14 / 57 | 24.6% |
| 39 | Wuhan Gelanruo Intelligent Technology Co., Ltd. | 3 | 12 / 48 | 0 / 3 | 1 / 3 | 0 / 3 | 13 / 57 | 22.8% |
| 40 | Sanctuary AI | 3 | 11 / 48 | 0 / 3 | 0 / 3 | 0 / 3 | 11 / 57 | 19.3% |
| 41 | ZTE Corporation | 3 | 5 / 48 | 0 / 3 | 0 / 3 | 0 / 3 | 5 / 57 | 8.8% |
| 42 | Wuhan Guanggu Huahui Technology | 3 | 3 / 48 | 0 / 3 | 0 / 3 | 0 / 3 | 3 / 57 | 5.3% |
Methodology and limits
The score is found slots ÷ (19 × catalog robot count). Each of the 16 spec fields counts once when its value has nonempty, resolving source IDs. Zero and false are valid values. All confidence categories count, including reported and estimated values; this measures coverage, not independent performance or evidence quality.
Price counts when a current price record has resolving source IDs, including targets, estimates and monthly subscriptions. An official buy path requires an order or sales-contact URL backed by at least one source classified as official. It does not imply stock, regional eligibility or general availability. A document counts when the curated datasheet link exists in a sourced robot record, or a robot-details document URL matches a listed source. Datasheet links in the base record lack separate field-level source IDs.
All generations, statuses and form factors are included and weighted equally. No adjustment is made for fields that may not apply to a torso or historical model. More complete research can raise a score without any maker changing its publications. Unknown values mean not found in the public sources we checked; our search depth, language coverage and verification dates differ. This is a current build snapshot, not a historical trend or a claim about a maker’s intent.
Report a missing source to improve the underlying records. Read the site methodology or download the catalog exports.