
BuyBest for most hard-floor homes
ECOVACS DEEBOT T90 PRO Omni
- Suits
- Hard-floor homes whose rugs are thin or mid-depth and whose pet's shedding is nothing special
- The catch
- Thick, plush rugs can stall it
Opens Amazon.com · Prices change often
Buying guides
Home-fit finder









Buying guide
Verdict
Swept's pick for wood, LVP and tile is the ECOVACS DEEBOT T90 PRO Omni: it lifts dried stains and leaves the floor nearly dry. Best with thin to medium rugs, since a thick rug can stall it.

BuyBest for most hard-floor homes
Opens Amazon.com · Prices change often
At a glance
SuitsHard-floor homes whose rugs are thin or mid-depth and whose pet's shedding is nothing special
The catchThick, plush rugs can stall it
SuitsSingle-level hard-floor homes with flush doorways, rugs that can't take a soaking, and pets that never slip up indoors
The catchDoesn't reliably avoid pet waste
SuitsHard-floor homes with deep rugs and a hard-shedding pet, where baseboards and toe kicks matter most
The catchLifts dried stains unevenly and leaves the most water of these picks
SuitsTile or sealed LVP with barely a rug down and no shedding dog, picked up before it runs
The catchCan't see cords, socks or toys
SuitsHomes that want nothing on the robot able to film, with a house-trained pet and someone who picks up first
The catchRuns over pet mess and cables it can't see
Why you see a price band, not a price Prices change daily, so our pages show only the price band a robot sits in and leave the current price to Amazon. That way a page is never wrong by Tuesday.
The cruel joke of the robot mop is that the surest way to hurt a wood floor is to leave it wet. Plenty of these robots do exactly that while promising a shine. Swept's pick for wood, LVP and tile is the ECOVACS DEEBOT T90 PRO Omni. In Vacuum Wars' stain test it lifted dried coffee and juice well above that lab's average. It also left about a third of the usual water behind. It belongs in a home whose rugs are thin or mid-depth and whose pet sheds nothing out of the ordinary. The catch: a thick, plush rug can stall it.
Half of it. Every box promises a cleaner floor, but the type of mop mostly predicts how wet that floor ends up.
Rugs are a separate question. The Hook Up sent seven flagship mop robots from hard floor onto carpet in the middle of a mopping run. Only one kept the rug dry: the Curv 2 Flow, whose roller lifts and hides behind a shield. The eufy Omni S2's roller has no cover, and it wet the carpet.
Some homes should look elsewhere. If a puppy still has accidents, a robot that sees mess beats one that scrubs it, so read our eufy Omni S2 review. For rooms at different heights and pets that never have accidents indoors, the Roborock Saros 20 is our pick for steps. Mostly carpeted homes should start with our carpet picks, and anyone who never wants a mop with our vacuum-only picks. At under $300, mopping means a dragged cloth. For a small hard-floor home held to that budget, our pick is the Tapo RV30 Max Plus, a vacuum first whose pad handles light spills only. If you can't tell which home you have, the home-fit finder will sort it out.
BuyBest for most hard-floor homes
Price band: under $600
ECOVACS's big promise is a roller rinsed with fresh water while it mops, so yesterday's dirt never gets smeared back across today's floor. On wood, that's the claim that matters, and it holds. In Vacuum Wars' dried-stain test the T90 earned 157, where that review's average was 108. The same lab weighed only 0.35 g of water left behind it. A three-month home test found the roller free of hair and the self-emptying strong.
It's good at the rest of the job too. It dodged 23 of 24 obstacles in the same lab, so the phone cable by the sofa isn't a death sentence. Pet hair flattened into that lab's carpet came up completely, and long hair never wrapped around its brush. The box quotes a 0.9 in (24 mm) climb, yet in testing it cleared a 1.3 in (32 mm) sill. A maker underselling its own robot is rare enough to frame. At 3.7 in tall with no turret, it slides under most sofas, and it can vacuum the whole floor before it starts mopping.
The roller reaches out about 0.6 in (1.5 cm) toward the baseboard, but the side brush doesn't extend. Nobody has tested how the US model copes with corners, so a kitchen full of cabinet corners is where we'd watch it first.
The catch is thick rugs. ECOVACS says the roller lifts 0.6 in (15 mm) on carpet, but nobody has tested whether that keeps a deep rug dry. Even ECOVACS's own sponsored three-month video admits it sometimes jams on a thick rug. Thin and medium rugs are fine. The shag in the den is not its friend.
The evidence is one lab, a three-month home test and owners. No second lab has weighed its mopping yet, so know it rests on one lab's numbers.
If your floors are wood, LVP or tile with thin to medium rugs, we'd buy it. Our full T90 review has the rest.
BuyBest if rugs sit on your hard floors
Price band: under $700
Roborock says that on carpet the roller lifts 0.6 in (15 mm) and a shield drops over it. That kind of feature usually works only in the brochure, but this one held. In The Hook Up's seven-robot run onto carpet, it was the only robot that left the rug dry. Its patch of dried mud also came up clean enough to pass a paper-towel check.
Then Vacuum Wars' dried-stain test gave it 25, in a review where the average was 112. Both results are real. That lab measured just 0.2 g of water left behind, under a fifth of its average. Tom's Guide saw streaks at the default water setting that vanished when it turned the flow up. Out of the box it mops with a light hand, and the testers and owners who ran it wetter got the clean floors.
It reaches edges by sliding the roller out. Roborock says that gets within about 0.4 in (10 mm) of the wall; nobody has measured it. It has no mop-only mode, and at 4.7 in tall it won't go under a low sofa. It climbs only about 0.75 in (19 mm), so a raised doorway stops it cold.
The catch is pet waste: its camera doesn't reliably spot it. In The Hook Up's test it rolled across the fake dog mess more than once.
It suits a single-level home with flush doorways and rugs that can't take a soaking. If your pets never slip up indoors, we'd buy it.
BuyBest for edges and deep rugs
Price band: under $1,700
The X60's box boasts of mop pads washed at 212 °F (100 °C). I spent years being handed readings taken at the most flattering point in a system. This is the same trick: Dreame's own fine print puts that temperature at the heater, not the mop. RTINGS probed the wash area and drain and read 77–86 °F (25–30 °C). Another tester found the hottest spot on a freshly washed pad at 94 °F. The kettle is real. It just never meets your floor.
What the X60 does brilliantly is reach. The Hook Up found its swing-out pad mopped up to the edges far better than the eufy S2's roller managed. Baseboards and the toe kicks under cabinets get scrubbed instead of skipped. Dreame says vacuum-only runs leave the pads at the dock, which is how deep rugs stay dry; on a mixed run it wet the rug like the rest. It gets over a 2 in (51 mm) sill and slips under furniture as low as 3.1 in. RTINGS found it reliably avoids pet waste.
The catch is the mopping itself. Vacuum Wars' coffee-stain test gave it 136, where that review's average stood at 93. Yet RTINGS called it sub-par on dried stains in every configuration. On The Hook Up's dried mud it cleaned little beyond the seams of the planks. It also left 1.4 g of water. A superb robot that mops like a good one on its better days.
It's for deep rugs on hard floors, a dog that sheds hard and baseboards you're tired of wiping. For that home, we'd buy it. Our X60 review covers its carpet and pet-hair record.
BuyBest mopping for less
Price band: under $400
Roborock's page waves a certificate for "0% hair tangling". Read the small print and it covers the side brush, which is like boasting your spare tire has never had a flat. The main brush wrapped 62% of Vacuum Wars' long test hair. Only 43% of the pet hair pressed into that lab's carpet came up, about half its average. Not for a home with a shedding dog.
On a hard floor it's a different machine. Its spinning pads scored 149 for dried stains, with 108 as that review's average. It left 1.15 g of water, a little above that lab's norm. That's fine on tile and sealed LVP, and worth watching on older wood. Roborock rates its wash water at 167 °F, and to its credit a footnote says that's measured at the heater. The pads don't swing out, so a three-month home test found edges and corners left behind.
The catch is that it's blind to small things. It avoided 5 of 24 obstacles, and the home test watched it run over a cable and toys. It needs a floor someone has picked up.
For tile or sealed LVP with barely a rug down and no shedding dog, we'd buy it.
BuyBest if you don't want a camera
Price band: under $500
eufy says the C28 "intelligently avoids obstacles". It tries with a line laser and no camera. The result sees the outline of your living room but not what's lying in it. It scored about average on Vacuum Wars' obstacle course, and in a three-month home trial it drove straight across fake dog poop and cables alike. For some homes that's the point. There's no camera or microphone on board, so nothing in the bedroom gets filmed.
Its roller is the quiet surprise of this page. It left 0.25 g of water, about a quarter of Vacuum Wars' average. It earned 119 in the dried-stain test, a little over the 108 its review gives as average. Wirecutter found it fine for daily mopping but nothing special on deep-set stains, and a step behind eufy's E25 at edges. Of the pet hair Vacuum Wars pressed into carpet, not a strand was left. eufy's user guide sets a limit of 0.7 in (18 mm) for rug thickness and asks you to tuck tasseled edges under. There's no mop-only mode.
The catch is the one you've chosen. It can't see mess, so it's for house-trained pets and floors someone clears first.
If nothing on your robot should be able to film your home, we'd buy it.
Rollers are this year's fashion. Roborock's first, the Curv 2 Flow, arrived in January and eufy's S2 in March. Both prove this page's point: a roller keeps the floor dry, and stains are a separate question.
The one we're watching is the MOVA V70 Ultra Complete, on sale since August. MOVA claims a mop that reaches out 6.7 in (17 cm) for corners. It also claims pads that stay at the dock for vacuum runs, which a lab confirmed. Vacuum Wars' dried-stain test gave it 133, above the 108 average in that review, and watched it mount a 2 in (51 mm) threshold. But that lab found its water output erratic, with more at the lowest setting than the highest. Owners report pads coming out too dry, and MOVA has talked of a firmware fix. Its 212 °F (100 °C) wash is a claim nobody has measured. The bagless dock is unproven with a heavy shedder, too. In one home with several shedding dogs, hair caked the dock's filter within a week. One lab and a few weeks of owners is too thin to buy on. Wait for the fix and a second test on a US unit.
We set each maker's mop claims against what labs and owners found on the exact US model. Water left on the floor counted most, then dried stains. Dry rugs and reach into edges settled the close calls. Every pick has at least two separate kinds of evidence on its exact model, one from outside the maker.
On sealed wood, with care. Consumer Reports notes that the National Wood Flooring Association warns against wet mops on wood. Moisture can damage the finish and the wood over time, and CR's testers saw robots leave some behind. Its advice is to check your flooring maker's guidance first. That's why water left on the floor ranks first here, and why drag pads fell away. No lab has tested robots for scratching. The owner reports we found blame trapped grit under particular models, not robots in general.
Nobody has shown that it does. Dreame's 212 °F became lukewarm by the time RTINGS measured it at the pad. Roborock's own footnotes say its 167 °F is measured at the heater. ECOVACS quotes 167 °F (75 °C) for the T90 without saying where, and nobody has checked it. Nobody has tested whether a hot wash keeps a dock from smelling, either. Choose on dried-stain results, which labs do measure. Hot water is a nice line on the box and nothing more.
Because they test different messes at different settings. Vacuum Wars scores dried coffee and grape-juice stains at two water settings and averages them. The Hook Up dries mud onto LVP and mops twice at full water. RTINGS tries several settings, up to the wettest. A light-handed mopper like the Curv 2 Flow can look weak in one test and excellent in the next. So we lean on results that more than one lab reproduced.
| Pick | Mop | Dried stains (Vacuum Wars score, that review's average) | Water left on the floor | Amazon |
|---|---|---|---|---|
| ECOVACS DEEBOT T90 PRO Omni | Self-rinsing roller | 157 (average 108) | 0.35 g, about a third of average | View on Amazon (opens in a new tab) |
| Roborock Qrevo Curv 2 Flow | Self-rinsing roller with rug shield | 25 (average 112) | 0.2 g, under a fifth of average | View on Amazon (opens in a new tab) |
| Dreame X60 Max Ultra Complete | Two spinning pads, one swings out | 136 on coffee stains (average 93) | 1.4 g, the most of these picks | View on Amazon (opens in a new tab) |
| Roborock Qrevo S Pro | Two spinning pads | 149 (average 108) | 1.15 g, a little above average | View on Amazon (opens in a new tab) |
| eufy Omni C28 | Self-rinsing roller | 119 (average 108) | 0.25 g, about a quarter of average | View on Amazon (opens in a new tab) |

BuyBest for most hard-floor homes
Plenty of robots promise a shine. The T90 leaves a floor dry enough to walk on in socks, and it spots the sock before eating it. It costs less than most of the robots it outscrubs. Keeping it off the deep shag is the only care it asks. For hard floors with thin to medium rugs, we'd buy it.
Research behind this page: 18 models weighed; 17 maker manuals among them; across this site's research, 497 sources kept in full; 148 Reddit threads read, kept as the reader's summary, not in full.
Vacuum Wars' dried-stain scores sit on that lab's own scale, and its average moved between reviews, so each score is shown with the average its own review gave. Water is grams left on a paper towel after wiping the area behind the mop.
| Robot | Mop | Dried stains (Vacuum Wars, that review's average) | Water left (Vacuum Wars) | Rugs and edges | Call here |
|---|---|---|---|---|---|
| ECOVACS DEEBOT T90 PRO Omni | Self-rinsing roller | 157 (108) | 0.35 g, about a third of average | Rug crossing not tested; roller extends, side brush doesn't; US-model corners not tested | Buy, overall pick |
| Roborock Qrevo Curv 2 Flow | Roller with shield | 25 at medium water, 38 at max (112) | 0.2 g, under a fifth of average | Only one of seven to keep carpet dry (The Hook Up) | Buy |
| eufy Omni C28 | Self-rinsing roller | 119 (108) | 0.25 g, about a quarter of average | Edges a tick behind the E25 (Wirecutter); no long hair wrapped at Vacuum Wars | Buy |
| eufy Omni S2 | Roller, no cover | 82 (108) | 0.35 g | Wet carpet on a mixed run; edge mopping 2 of 10 (The Hook Up) | Not picked here; see its review |
| eufy Omni E25 | Self-rinsing roller | 115 (110) | 1.1 g, a touch above average | No roller cover; dried coffee not lifted by two testers | Not picked here |
| Matic | Microfiber roller | 147 (average not stated) | Not reported | RTINGS: treats stains as obstacles, misses corners | Not picked |
| Dreame X60 Max Ultra Complete | Two spinning pads, one swings out | 136 on coffee (93) | 1.4 g | Edge mopping 7 of 10 (The Hook Up); pads stay at dock for vacuum runs | Buy |
| Roborock Qrevo S Pro | Two spinning pads | 149 (108) | 1.15 g, a little above average | Pads don't extend | Buy |
| Roborock Saros 20 | Two spinning pads, one extends | 116 (94) | 0.8 g, a little under average | Edge mopping 6 of 10; pads dragged on carpet, can stay at dock | Not picked here |
| Roborock Saros 10R | Two spinning pads, one extends | 103 (above that review's average) | Slightly above average | Pads detach for vacuum runs | Not picked; its review says wait |
| Dreame L60 Ultra PE | Two spinning pads, one extends | 85 (108) | 0.3 g, under a third of average | Climbs about 0.75 in | Not picked here |
| MOVA V70 Ultra Complete | Two spinning pads, one extends | 133 (108) | Little, but erratic | Pads stay at dock for vacuum runs; bagless dock unproven with heavy shedders | Wait |
| MOVA P10 Pro Ultra | Two spinning pads, one extends | 73 (93) | 1.05 g, about average | No side brush on carpet (RTINGS) | Not picked here |
| Dreame L40 Ultra Gen 2 | Two spinning pads, one extends | Well above average (no figure published) | About average | No pad drop; washes in room-temperature water | Too early to call |
| Narwal Freo X10 Pro | Two spinning pads, one extends | Not on this scale | Not reported | Wirecutter: spreads puddles, tracks wet mud; strong at edges | Not picked |
| Roborock Q7 M5+ | Drag pad, no lift | 158 (93) | 2.0 g, twice average, at every setting | No lift; carpets must be marked in the app | Not for mopping wood |
| Dreame D20 Plus | Drag pad, no lift or wash | 90 (107.8) | Not reported | Pad comes off for carpet | Not picked |
| Tapo RV30 Max Plus | Clip-on pad, no lift | Not tested by Vacuum Wars | Not reported | Tom's Guide: light spills only, can leave residue | Not picked |
| Robot | Why it isn't picked here | Read instead |
|---|---|---|
| eufy Omni S2 | Best here at seeing pet mess, but weak on dried stains at two labs, and its 1.1 in (28 mm) mop lift with no cover wet the rug on The Hook Up's mixed run | eufy Omni S2 review, for a puppy that still has accidents |
| Roborock Saros 20 | Labs found its mopping ordinary to underwhelming, and it can't be trusted around pet waste | Saros 20 review, for steps between rooms |
| Roborock Saros 10R | Strong stains at RTINGS and excellent avoidance, but contested pet hair | Saros 10R review, which explains the wait |
| eufy Omni E25 | More water than the other rollers, and two testers saw dried coffee left behind | Picks under $600, for homes with lots of rugs |
| Dreame L60 Ultra PE | A carpet and pet-hair robot whose mop runs too dry to lift dried stains well | L60 Ultra PE review |
| Dreame L40 Ultra Gen 2 | New this year; good lab stain scores, but a three-month home test called its mop poor, and in its band the T90 does what it can't | Nothing yet |
| MOVA P10 Pro Ultra | Below its review's average on dried stains; its strength is seeing mess for less | Picks under $600 |
| Roborock Q7 M5+ | Strong on stains, but its drag pad leaves the most water of anything here | Q7 M5+ review, for small, tidy apartments |
| Dreame D20 Plus | A vacuum first, with a cloth that drags | D20 Plus review, our under-$300 pick for carpet, rugs and shedding pets |
| Tapo RV30 Max Plus | A vacuum first; its clip-on pad handles light spills only | Tapo review, our under-$300 pick for a small hard-floor home |
| Narwal Freo X10 Pro | Good at edges, but its spinning pads spread puddles and it has no lab stain score on this scale | Nothing here |
| Matic | Two labs split badly on its mopping, and RTINGS found it steered around stains instead of cleaning them | Nothing here |
On the Curv 2 Flow, Vacuum Wars and Notebookcheck found it weak on dried stains, while The Hook Up's dried mud came off well enough that its floor passed a paper-towel test. The likely causes are water setting (Vacuum Wars measured very little water; Tom's Guide found the default too low), stain type (dried coffee against mud on LVP) and number of passes. Owners running high water side with The Hook Up. Its pick rests on keeping rugs dry, which no lab disputes.
On the X60, Vacuum Wars scored its coffee-stain test above that review's average; RTINGS (bought unit, firmware 4.3.9_1495, reviewed July 2026) called dried stains sub-par in every configuration; and The Hook Up saw it clean only the seams of its dried mud. Jamie Andrews saw baked-on hot sauce go in one pass at deep mode and max water. RTINGS measured 77–86 °F (25–30 °C) wash water in High-Temperature Soaking mode; Jamie Andrews found 94 °F at the pad.
On the S2, Vacuum Wars and RTINGS found it weak on dried stains, while Just A Dad Approved rated it well after firmware improved dirty-water capture. Firmware arrives every week or two, so labs that tested months apart may have tested different software.
On the V70, one exact-model owner with several heavy-shedding dogs found hair collecting at the dock's suction port and in its pipe, a bagless tank filling faster than the old V50's bag, and a caked dock filter after a week. Another owner in the same thread, whose variant isn't confirmed, has two long-haired German Shepherds and no such trouble. Vacuum Wars, the only lab so far, didn't see the dry-pad problem in its studio.
On the T90, Reddit threads whose owners never confirmed the US model describe fine streaks, a slipping brush-drive gear and detergent dosing that stopped. We don't count them as T90 evidence, but they're worth watching. A weaker corner clean was reported on a unit from another market, which we don't count either; nobody has tested the US model's corners. ECOVACS claims 90 days per dust bag; its sponsored three-month video got about 2 months with cats, a dog and litter.
Vacuum Wars buys its units. Its stain test uses dried coffee and grape-juice stains run at two water settings and averaged, and its water penalty weighs a paper towel wiped behind the mop. It reviewed the T90, Curv 2 Flow, X60, Qrevo S Pro and C28 on US units, and the V70 in an October 2026 video.
The Hook Up compared seven 2026 flagships, unsponsored, in April 2026, with dried mud on LVP mopped twice at max water, a mid-mop carpet crossing, and edge and corner tests. The T90, Qrevo S Pro and C28 weren't in it. RTINGS bought the X60, Curv 2 Flow and S2 in the US. Tom's Guide reviewed the Curv 2 Flow in February 2026, and Wirecutter's mop-combo guide covers the C28, E25, D20 Plus and Narwal (August 2026).
The longer home tests came from Just A Dad Approved, who ran the T90, Qrevo S Pro and C28 for about three months each, and from Daniel Romero, whose three-month T90 video is sponsored by ECOVACS; Jamie Andrews lived with the X60 for about a month. Consumer Reports supplied the wood-floor guidance, quoting the National Wood Flooring Association.