How we score
Every robot gets an overall score out of 100 plus a breakdown by criterion. Our scores are independent: no advertiser can buy or sway them. Here is exactly how it works, starting with humanoids.
1. We gather the data
For each robot we collect the maker's published specs, press-verified figures and, where it is sold, real prices from several merchants. Every humanoid figure is sourced; if a maker hasn't published a number, we leave it blank rather than estimate it.
2. We score each criterion
Specs are converted to 0–100 scores using scales that are the same for every model in a category: degrees of freedom, hand dexterity, payload, battery life, top speed and availability for a humanoid; suction, navigation, battery and value for a robot vacuum. That keeps the scores comparable.
3. How humanoids are scored
A humanoid's overall score is the average of its six criteria. Height is shown but not scored (taller is not better). A spec the maker hasn't published counts as 50, because a figure you can't check is one you can't plan around. Availability scores 100 for a robot that ships today, 84 on pre-order, 70 in paid pilots, 52 for a prototype and 40 once discontinued. Prices announced in dollars, yuan or pounds are converted to euros at an approximate rate, and the original figure is shown on the product page.
4. We compute the overall score
The overall score sums up how the robot performs within its category. A model is never judged against a different category: a humanoid and a quadruped don't share criteria.
5. We keep it current
As prices move and new models launch, scores and rankings recalculate. Our best-of guides reflect the market as it is now, not a frozen snapshot.
Our score scale
We earn an affiliate commission if you buy through our links, but it never influences a score or ranking. See our affiliate disclosure for the details.
Affiliate disclosure →