

Table of Contents
I'm asked all the time what robot vacuum to buy and most of the time I tell people to go with a Roborock. I've been impressed with most models they've released over the last few years. Through many iterations, they consistently vacuum and mop with the least amount of fuss and the most efficacy. But this month, I was bowled over by the S10 Ultra from newcomer 3i. These two models share lots of functionality, with only minor differences between them. While both would be great, reliable choices, you can make your choice easily, based on which is more important to you: form or function?
Sleek and subtle vs. large and showy

Robot vacuum towers are a pain worth living with—they're large but spare you the manual labor of emptying and cleaning out your robot every day or two. That said, some towers make the living easier, and on this point, Roborock has the upper hand. While Roborock towers are always sleek and modern, the Saros 10R tower is a genuinely beautiful piece of work. With a clean, mirrored face and metal accents, the tower can be tucked between furniture or under a coffee table.
By comparison, the 3i S10 tower is pretty showy. Weighing 50 pounds without water, cleaner or a robot in it, it’s massive, with a futuristic design that literally glows. What it lacks in subtlety, however, it makes up for in function.
While the Roborock has decent-sized clean and dirty water tanks on board, the 3i doesn’t have a dirty water tank at all: it recycles water on board through a distillation process, and generates water from the air via dehumidifying. Both towers have a slot on board for cleaning fluid, and both did an excellent job suctioning out all dry debris into the vacuum bag and cleaning the mop pads with hot water.
While I’d personally favor never having to change the water in the robot tower, this truly is form vs. function. You want to like the aesthetics of the things in hyour home, and the S10 is … a lot to live with. I’d understand someone favoring the Roborock tower. Let’s call it a wash.
Robot apps are great when features are easy to access

I’ve installed almost 40 robot vacuums over the years, and there’s almost nothing different about how they get installed. While I thought the directions on the 3i were not great and the illustrations led me to install the cleaner upside down, ultimately, both robots were installed and paired in the app in under 10 minutes.
Where these robots differ is in the app, where they share many functions, buried in tabs, accordions, submenus and pages. It would be fair to say each app has a learning curve, and I’d likely learn the location of functions in the 3i app eventually. Even so, I think Roborock is a cleaner experience. Things felt unnecessarily buried in the 3i app. I noted in my review that I was once able to find the button to tell the tower to empty the robot a second time, but was never able to find it again. That’s because the button is only available when the robot is docked. I’m sure an engineer thought this made perfect sense, but it’s confusing and frustrating to not see all your options in routine, recognizable places.
Roborock has arranged the functions in more user-friendly places, but they’ve also produced a lot more robots and have years of user experience to rely on. The 10R has two features I find indispensable: Pin and Go, and Remote Control. Pin and Go lets you set a pin on the map, the robot immediately goes there in the most direct route, and cleans around the pin. You don’t have to set a zone (which are never very accurate). Remote control allows you to direct the robot from wherever it is; the best use is to rescue your robot when it’s stuck under your couch or other furniture, so you don’t have to get on the floor and try to fish it out. The S10 has neither of these features, and I missed them.
The 3i app is perfectly usable; there’s nothing about the experience that would turn you off from the S10 altogether. But side by side, I favor the Roborock experience. The function commands are closer to the surface in the app.
Both robots are adventurous
The Saros 10R has a huge distinction from other modern robots: It doesn't use LiDAR (a common robot navigation technology that measures light distance) at all. It relies solely on cameras and AI. The S10 Ultra uses both. Since the 10R is the only robot I've tested without LiDAR, it's hard to know what is unique to losing LiDAR vs just being a feature of the 10R. What impressed me about the 10R, immediately, was how quickly it mapped the entire house (it took only seconds) and how the robot seemed willing to get into tighter spaces than previous robots were. I naively suggested this might be due to the AI being great, and LiDAR being less great.
I was delighted when the S10, with LiDAR and AI on board, went into the same exact spaces. Again, previous robots weren't willing to try and fit into tight, dark spaces, but both robots here were.
There's no contest when it comes to vacuuming power

Where the rubber (or robot) meets the road is the vacuuming power. In the realm of vacuuming, I look for three kinds of debris removal: microdebris like dust; medium-sized debris like rice or cereal; and macro debris like pet toy fluff or clothing tags. Most people have a reasonable amount of dust and small detritus on the floor, but parents of humans and animals alike will have a lot of macro debris as well. This is the stuff that challenges robots (all vacuums, actually) the most. Debris may get stuck in the roller, or the tower may have a hard time suctioning it out of the robot.
While the 10R performed to the standard I expect from Roborock premium robots, the 3i handily surpassed every robot vacuum I have ever tested, and did it conclusively. There was almost no debris, no matter how large or small, that didn’t get caught by the twin floor sweeps, which deliver detritus efficiently to the robot roller. Large pieces of mulch (like small twigs), paper receipts, hair ties: The 3i S10 ate them all up without so much as a soft rattle and went on its way.
Again, this isn’t to say that the 10R is a dud. It performed fine, but not as well as the 3i. The 10R has the same frailties as most robot vacuums, where AI tells it to go around dog toy fluff. A piece of mulch will get caught in the rollers and bring everything to a halt. The new split roller design of the Saros line definitely helped the 10R get caught on less than say, the Roborock Curv. But it still left a lot of large debris on the floor, and when it tried to suck it up, it often got stuck in the rollers, requiring human intervention. The extending arm of the 10R does help get debris out from corner of walls, from under thresholds and whatever furniture sits on the ground. The 10R has a fantastically high suction power, at 20,000 Pa, and the vacuum will very competently suck whatever it rolls over off the floor.
Still, this wasn’t a close call. The 3i is a ridiculously capable vacuum, and if you’re looking for the vacuum that you won’t need to sweep ahead of time for, this is it.
Different approaches with the same result

The 10R and S10 take completely different approaches to washing the floor. The 10R uses the popular twin spinning mop heads on the back of the robot; the S10 Ultra uses a singular, long mop brush. Both extend from under the robot to reach walls, although the twin mop heads on the 10R are certainly more flexible. They can extend on three sides, whereas the singular mop brush on the S10 Ultra extends only on the right side, as the robot rolls alongside a wall. I’d like to say this difference matters, but it didn’t, in my experience.
The 3i has a slight advantage here in hardware: Many jets of water clean the mop brush constantly, even while working on the ground. The Roborock has no such functionality—it gets a deep, hot clean when it returns to the dock. I’d like to say this had an effect, but it was purely theoretical. I saw no difference in how clean the floor got.
For both mops, the results were largely the same: Any wet surface liquids were easily cleaned up. Most stains were picked up, by either robot. They both tackled muddy paw prints with aplomb. Both were able to get close to walls. However, neither robot was able to get the truly ground-in grime off the tile floor. That’s not to either robots' discredit; I’ve yet to encounter a robot mop that could. In either case, i found the tower did a great job cleaning the mop after service. Both mops emerged looking fluffy, clean and new.
Both robots did a completely acceptable “everyday” mop job on the floor, regardless of whether it was tile or hardwood. You’re still going to need to keep a mop around to dig in on ground in grime, but either robot will handle a lot of light work for you.
And the Winner Is...

While I could easily make an argument for either robot vacuum, I'm all about function in my home. I just want the damn floors clean, despite the constant assault on them by dust, dogs and my own penchant for dropping things. For this, I will put up with a mildly obtrusive looking tower.
While the Saros 10R currently retails for $1599.99 and the S10 Ultra for slightly less at $1399.99 (on sale for Amazon Prime members, down from $1899.99, both models are frequently on sale and may flip-flop on price over the coming months. Currenly at $1399.99, I think the S10 Ultra is a better value.