L10s Ultra Gen 3
2 independent reviewsRead the findings →A mid-range Dreame that empties itself and washes and dries its own mops. It suits homes with mostly hard floors and shorter rugs. Long hair can wrap its brush, though, and it's inconsistent at dodging small objects.
Updated
See the alternatives below, or switch market.
Feedback
What reviewers found
Two hands-on tests, both from Germany and both positive. Computer Bild put it top of its list of budget robot vacuum-mops, with strong suction, thorough mopping and a comfortable dock. Notebookcheck called it a lot of robot for the money, especially on carpet. They split sharply on obstacles: Computer Bild saw it reliably avoid cables, socks and even dog mess, while Notebookcheck watched it run over small toys and shoelaces. Both flag that it struggles with raised doorway thresholds and has no hot-water wash.
Computer Bild, hands-on, Jan 2026 · Notebookcheck, hands-on, Dec 2025
Hard floors
Reviewers rate it highlyComputer Bild saw it collect almost all of a cat litter, muesli and wood-shaving mix and clean edges well if not perfectly, and Notebookcheck found the extra suction paid off on smooth floors.
Carpet and rugs
Reviewers rate it highlyNotebookcheck singles out carpet as a strength, with short pile cleaned well and only a few grains of sand left in a longer-pile rug. Computer Bild found it thorough on carpet, boosting suction automatically.
Pet hair
Not enough evidenceOnly Notebookcheck covered hair, finding long hair tangled round the plain rubber brush more often than it would like, while short dog hair caused no trouble.
Mopping
Mixed resultsGood for upkeep, not a deep scrub. Computer Bild found it mopped evenly right to the edges, but stubborn stains needed several passes and streaks sometimes remained, and Notebookcheck found greasy dirt beyond it.
Navigation and obstacles
Mixed resultsThe testers split on small objects. Computer Bild saw it reliably avoid cables, socks and even dog mess, while Notebookcheck watched it catch building bricks and shoelaces. Both found doorway thresholds over about 2cm too much.
Dock and upkeep
Reviewers rate it highlyBoth testers liked the dock, which empties the bin, washes the pads and dries them with warm air. Computer Bild found it easy to clean and narrow, if tall, and Notebookcheck found the mops generally came out clean. There's no hot-water wash.
App and software
Reviewers rate it highlyNotebookcheck found setup trouble-free and the app full of options, including Matter support, and Computer Bild found it clearly laid out, with flexible schedules, room order and suction settings.
Noise
Not enough evidenceNone of the reviews we use measured or judged its noise.
Reliability and support
Not enough evidenceNone of the reviews we use ran it long term or judged Dreame's support.
Value
Reviewers rate it highlyNotebookcheck calls it a lot of robot for little money, and Computer Bild rates it one of the best robot vacuum-mops in its price bracket.
Verdicts come from independent, hands-on, finished reviews. An area needs at least two of them before we say more than “Not enough evidence”. We don't score robots ourselves. How this works
Testers disagree on obstacles
Computer Bild found it recognised cables, socks, shoes and even dog mess, and steered round them reliably. Notebookcheck had the opposite result with smaller items: building bricks, shoelaces and similar objects were caught one after another. Its single-laser sensor handled chair legs and larger objects fine in both tests. The difference may come down to object size, but nobody has shown it. Pick small things up before it runs.
Strong suction, especially on carpet
In Computer Bild's test it collected 97.4g of a 100g mix of cat litter, muesli and wood shavings, and cleaned edges well if not perfectly. Notebookcheck found the extra suction paid off on smooth floors and short-pile carpet, though a few grains of sand stayed in a longer-pile rug on maximum power. For everyday dirt on typical floors, this is the reason to buy it.
Long hair wraps the brush
Notebookcheck found long hair tangled around the main brush more often than it would like. As its comparison with the Gen 2 points out, Dreame has swapped the older model's anti-tangle brush for a plain rubber roller, so this is a step backwards. Short dog hair caused no trouble. Computer Bild didn't mention tangling either way. If anyone at home has long hair, expect to cut strands off the brush regularly.
Thorough mopping, but repeat passes
Computer Bild found it mopped hard floors evenly and right up to the edges, but stubborn stains only came off reliably after several passes, and the finish wasn't always streak-free. Notebookcheck rated the mopping fine for regular upkeep but found greasy or stubborn dirt too much for it, though it welcomed the automatic detergent dosing. Expect well-kept kitchen floors, not a deep scrub.
A capable dock, without hot water
Both Computer Bild and Notebookcheck describe a station that empties the robot, washes the pads without hot water and dries them with warm air. Computer Bild found it easy to clean and pleasingly narrow but tall at nearly 59cm. Notebookcheck found the mops generally came out clean, only struggling with coarser dirt, and notes an optional kit that plumbs it into your water supply. Check the dock's height against where you want it to live.
These tests cover the L10s Ultra Gen 3 only, not the Gen 2 or the original L10s Ultra.
Our take
Should you buy the L10s Ultra Gen 3?
Best for: Homes with mostly hard floors and a few rugs. Skip it if: Long hair is a frequent problem, or your rugs are thick enough to touch the raised mop pads.
The Gen 3 is a reasonable option for mainly hard floors and shorter rugs, with a dock that handles emptying and pad washing. Long hair can wrap around its brush, and obstacle avoidance was inconsistent. A model with detachable pads may suit thick rugs better, but check hair pickup and obstacle results separately before choosing an alternative.
The Gen 3 adds a higher advertised suction figure and a self-cleaning washboard. Its dock empties dust, washes pads in unheated water, dries them with warm air, refills the robot and adds detergent. Matter support is intended to connect it with compatible smart-home systems; available controls depend on the system and software. It uses a rubber brush and laser-based obstacle sensing, without a camera.
Computer Bild collected nearly all of its hard-floor debris mix, and Notebookcheck reported strong carpet cleaning. Small-object results differed: Computer Bild avoided cables, socks and simulated pet waste, while Notebookcheck caught bricks and shoelaces. Long hair also tangled around the brush in Notebookcheck’s test. Stubborn stains needed repeated passes, and greasy dirt remained a weakness. Threshold results should be read separately: Computer Bild reported difficulty above about 1.5cm, while the other review described a limit nearer 2cm. Neither supports a general claim that every 2cm threshold is manageable.
Before you buy
The questions people ask about this one
Will it avoid cables and pet mess?
The tests disagreed. Computer Bild saw it avoid cables, socks and simulated pet waste; Notebookcheck had it catch small toys and shoelaces again and again. A few good runs don't make it safe around pet accidents, so clear small items and check the floor before each clean.
Can it cross rugs and raised thresholds between rooms?
Not well above 2cm. Computer Bild found it crossed thresholds up to 1.5cm easily but couldn't climb higher steps, and Notebookcheck also warns about doorway thresholds over 2cm. Rug edges were no problem for Computer Bild. If you have raised thresholds, look at a model with climbing legs.
Can it leave the mops in the dock for carpet?
No. The mops lift 10.5mm on carpet, per the specifications, but can't be left behind. Notebookcheck suggests running the carpet vacuuming first and then mopping the remaining rooms. On deep carpet, set no-mop zones in the app.
Does it fit under low furniture?
It's 9.7cm tall, per the specifications, and Notebookcheck points out the laser turret doesn't retract, so it won't reach under furniture as low as some pricier rivals. Measure the gap under your sofa and bed first.
Is it good with long hair?
Not great. Notebookcheck found long hair tangled on the rubber main brush more often, because this generation drops the anti-tangle brush the Gen 2 had. Short dog hair was fine. Budget a few minutes a week for cutting hair off the brush.
Should I get the Gen 3 or the Gen 2?
Notebookcheck reports that the Gen 3 adds a higher advertised suction figure, a self-cleaning washboard and Matter support. Which brush comes in the box can vary by bundle, so check the seller’s exact contents. Brush tangling has been reported on both generations, so don't pick the older one just to avoid cutting out long hair.
Alternatives
Other robot vacuums to compare.
L10s Ultra Gen 2
Why consider it: Home tests found useful everyday cleaning and a convenient dock.
The compromise: Missed edges, scattered debris and odd routes mean checking its first few runs.
D30 Ultra
Why consider it: Two tests found strong vacuuming and a long range on each charge.
The compromise: It's poor at dodging clutter, scrubbing stains and lifting flattened pet hair.
Omni C28
Why consider it: Tests agree on useful vacuuming and roller mopping.
The compromise: Obstacle handling and awkward furniture layouts are the clearest compromises.
Specs
Full specifications
Each specification names its source. Manufacturer claims are not independent test results; unconfirmed figures are labelled.
Owner reviews
Do you own the L10s Ultra Gen 3?
Tell people how it’s held up: cleaning, maintenance, faults and support. Owner accounts are shown separately from independent reviews.
Sources
Where this page’s findings come from
Editorial reviews
Sources
Videos
- Critical Tech: Dreame L10s Ultra Gen 3 | The Robot Vacuum That Does EVERYTHING?! ↗
- dBeta: ¿El TOP VENTAS del 2026? Dreame L10s Ultra Gen 3 Kit, Review! ↗
- Search YouTube for Dreame L10s Ultra Gen 3 ↗Search results may include company demonstrations or similar models. Check the source and model name; a demonstration is not an independent performance test.