Launched in October 2024 at a price of €1099, the Dreame Matrix 10 Ultra robot vacuum cleaner is positioned against established competitors such as the Roborock S8 MaxV Ultra (€1099), the Ecovacs Deebot X2 Omni (€1199), and the Dreame X50 Ultra Complete from the same brand (€799). The question of value for money is all the more relevant given the impressive performance figures: a suction power of 30,000 Pascals (a market record according to the manufacturer), 5.5-liter clean water tank and 4-liter dirty water tank, mop washing and drying at 100°C, and retractable Lidar navigation allowing it to go under furniture up to 8.9 cm high.
Our testing protocol spanned three weeks of daily use in a 120 m² house with tiled floors, varnished parquet flooring, and a medium-pile carpet. We subjected the appliance to standard usage scenarios as well as extreme dirt tests designed to verify its limits: spilling 200 grams of various debris (rice, pasta, lentils, crumbs), spreading and drying melted chocolate on the tiles, leaving ketchup stains for several hours, and intentionally flooding the bathroom with 5 liters of water. A thermal imaging camera measured the actual washing and drying temperatures, a sound level meter quantified the noise level, and a precision scale assessed the absorption capacity of each type of mop pad. Every claim made by the manufacturer was systematically compared to measurable reality.
Upon unboxing, the Dreame Matrix 10 Ultra arrives in a double-layered protective box, a sign of its premium positioning. The accessory package is generous: six replacement mop heads (two of each type), two side brushes, two HEPA filters, three 3.2-liter dust bags, one liter of standard detergent solution, two bottles of pet odor neutralizer, and two bottles of hardwood floor cleaner. This abundance allows for approximately twelve months of use without additional purchases. The docking station features a vertical design with a gold band running its entire height, giving it a luxurious look that contrasts with previous models from the brand. The materials used inspire confidence: thick plastics with no play or creaking, rigorous assembly, and meticulous finishing. The imposing dimensions (42 cm wide, 59 cm deep, 87 cm high with the door open) are due to the presence of the automatic mop head changing mechanism, a vertical system that requires space.
The central question remains: does this interchangeable mop system offer real, measurable added value, or is it simply a sophisticated marketing ploy to justify a €300 price premium over the X50 Ultra Complete from the same brand? Do the suction and washing performance figures live up to the advertised specifications? Do the washing and drying temperatures actually reach the claimed 100°C? Is the price-performance ratio justified in the face of fierce competition offering similar features? Here are the answers provided by our rigorous, unbiased, and impartial tests.
Dreame Matrix 10 Ultra
Robot vacuum cleaner and mop with automatic changing mop system. Record-breaking suction power of 30,000 Pascals, giant 5.5L and 4L tanks, hot mopping and automatic drying. Retractable Lidar navigation, anti-tangle brush, obstacle clearance up to 8 cm. 100% autonomous docking station with intelligent mop changing based on the room.
Design and build quality: rigorous assembly despite some reservations
Examination of the docking station reveals meticulous attention to detail in its construction. The plastics used are of a satisfactory thickness, estimated at approximately 3 mm for the main panels. No play is detectable in the joints, and the various components fit together with millimeter precision. The front door, which provides access to the mop compartment, opens smoothly thanks to a discreet magnetic button. The locking system inspires confidence with its two points of contact ensuring a tight seal. In use, after several hundred openings during our three-week test, no premature wear was observed.
The gold band running vertically the entire height of the station is a deliberate decorative element. Unlike some competitors who favor discretion, Dreame has opted for a bold design that won't go unnoticed in a kitchen or laundry room. This aesthetic choice will appeal to some and not to others, depending on individual tastes, but it demonstrates a desire to stand out visually. The embossed, gold-plated Dream logo reinforces this impression of a premium product. From a purely functional standpoint, these decorative elements add nothing, but they contribute to the psychological justification of a higher price.
The water tanks are a major consideration. The 5.5-liter clean water tank boasts the largest capacity ever seen on a robot vacuum. Its slim design with a narrow base minimizes its footprint while maximizing usable volume. A clearly marked maximum level prevents overflows during refilling. The 4-liter dirty water tank features a similar design. Both tanks are easily removed and reinserted thanks to ergonomic handles. However, their weight when full (approximately 6 kg for the clean water tank and 4.5 kg for the dirty water tank) could pose a problem for the elderly or those with reduced mobility. The lack of a drain valve on the dirty water tank necessitates carrying it to the sink or toilet, an inconvenient task given its weight and size.
The mop storage compartment, accessible by opening the front door, reveals a sophisticated mechanism. Six magnetic slots hold the three pairs of mops, arranged vertically according to a strict color code: gray at the bottom, blue in the middle, and orange at the top. These slots incorporate powerful magnets that automatically and precisely position the mops. The hook-and-loop system that secures the mops to their holder ensures a firm hold during rotations. At the bottom of the compartment, a vertical lift mechanism retrieves and places the mops onto the robot. This fully automated system works by magnetizing the mops, which then adhere to the mobile cart. During our tests, no recognition or positioning errors were observed, demonstrating a well-engineered mechanical design.
The robot itself has a diameter of 35 cm, standard for this type of device. Its adjustable height, from 11.1 cm (Lidar deployed) to 8.9 cm (Lidar retracted), is advantageous for cleaning under low furniture. The laser rangefinder retraction mechanism operates automatically, without user intervention. On the top, two physical buttons provide basic control: start/pause and return to charging station. They require firm pressure to press, preventing accidental activation. The top cover opens with a simple gesture to access the 310 ml dustbin. This bin is easily removed using a handle and includes a HEPA filter housed in a dedicated compartment.
An examination of the underside reveals the cleaning mechanism. The double central brush, the centerpiece of the suction system, features angled bristles designed to naturally deflect hair outwards to prevent tangling. This brush, mounted on a movable chassis that can be raised 5 mm off the floor in mopping mode only, boasts a robust construction with metal bearings at its ends. The extendable side brush, mounted on an articulated arm, can extend several centimeters from the chassis. Its automatic extension and retraction mechanism operates smoothly, with a flexibility that protects baseboards from potential impacts. The two rotating mops are mounted on a particularly ingenious articulated system: each mop has its own oscillating arm that maintains contact with the floor, even on uneven surfaces. This mechanical system, reminiscent of vehicle suspensions, represents a true feat of engineering.
The drive wheels, approximately 7 cm in diameter, feature grooved treads that ensure good traction on all types of surfaces. These wheels can be raised by 15 mm thanks to jacks integrated into the chassis, allowing them to overcome obstacles or better handle thick carpets. The front idler wheel, 3 cm in diameter, is mounted on a suspension system that absorbs shocks when traversing obstacles. This detail demonstrates a focus on protecting the floor and reducing vibrations.
One area for improvement concerns the cleaning base inside the station. Unlike the Roborock S8 MaxV Ultra or Ecovacs X2 Omni models, which offer a removable base for easier deep cleaning, the Matrix 10 Ultra's base remains fixed. If manual cleaning is required, the user must bend into the station in an uncomfortable position to reach the corners. Dreame justifies this choice by the absence of a solid debris collection grid and by a very large diameter drain hose (estimated at 25 mm) which prevents blockages. This technical design choice is understandable, but raises questions about long-term maintenance. Furthermore, two small tabs located inside the mop compartment door activate the safety contacts. These elements, apparently molded from thin plastic, seem fragile. Accidental breakage of one of these tabs would prevent the automatic changing system from functioning, a point worth noting.
Measured performance: contrasting results between vacuuming and washing
The Dreame Matrix 10 Ultra's key selling point is its 30,000 Pascal suction power. To verify this specification, we implemented a standardized test protocol. A full 200-gram vacuum cleaner bag was emptied onto 2 square meters of tiled flooring in the entryway: 50 grams of rice, 50 grams of uncooked pasta, 30 grams of lentils, 40 grams of breadcrumbs, and 30 grams of reconstituted dust bunnies. The robot was programmed to perform three passes in standard mode (the second of four levels), following a typical route that began by cleaning the edges and then tackled the center in parallel lines spaced 20 cm apart.
First pass result: weighing the dustbin contents revealed 170 grams vacuumed, representing 85% of the debris. Visual inspection confirmed the presence of a few scattered particles on the hard floor and a concentration of residue on the medium-pile carpet. Along the baseboard, a critical area, approximately 95% of the lentils were captured, a remarkable performance attributable to the extendable side brush, which deploys its arm 4 cm from the chassis to reach corners. Second pass with intensive carpet mode activated: weighing revealed an additional 25 grams vacuumed. Third pass: 5 more grams collected. Overall: 200 grams vacuumed in three passes, representing 100% of the initially collected debris. This result validates the manufacturer's claims regarding suction power.
We repeated this test in turbo mode (maximum level) with a single pass. Result: 185 grams vacuumed, representing 92.5% of the debris. The increased suction power partially compensates for the reduced number of passes, but doesn't allow for complete removal in a single pass. This observation suggests that a multi-pass strategy remains more effective than raw power, a useful lesson for configuring the device according to your priorities (time vs. results).
On a medium-pile carpet (12 mm high), we scattered 50 grams of a flour and powdered sugar mixture along with 20 previously collected long hairs. The robot, set to intensive carpet cleaning mode, made two passes. The result: 48 grams of the powder mixture were vacuumed up (96%), and 19 out of 20 hairs were recovered. The missing hair had been thrown outside the test area during the brush's passage. Examination of the central brush after the test revealed no tangled hairs, confirming the effectiveness of the anti-tangling system. This represents a significant time saving in maintenance, as traditional brushes require tedious weekly detangling.
The noise level was measured using a calibrated sound level meter placed 1 meter from the robot. Silent mode: 55 decibels. Standard mode: 60 decibels. Turbo mode: 68 decibels. Max mode: 72 decibels. These values are in the upper range for this segment, consistent with the advertised suction power. For comparison, the Roborock S8 MaxV Ultra registers 66 decibels in maximum mode, while the Ecovacs X2 Omni reaches 70 decibels. The Dreame Matrix 10 Ultra is therefore slightly louder, an acceptable compromise given its superior power.
Regarding floor cleaning, the results are more varied. We methodically tested each type of mop to evaluate their respective performance. First step: measuring physical characteristics. Protocol: each mop was soaked in a container of room temperature water (20°C) until completely saturated, then weighed after 30 seconds of draining. Gray mop (bathroom): 43 grams of water retained. Blue mop (living areas): 35 grams. Orange mop (kitchen): 37 grams. For comparison, a standard Dreame mop from the previous generation retained 33 grams. The gray mop therefore has a 30% higher absorption capacity than the old model and 23% higher than the current blue mop.
Testing the gray mop under real-world conditions: the bathroom (3.5 m²) was intentionally flooded with 5 liters of water, evenly distributed, creating puddles 2 to 5 mm thick. The robot was programmed in mopping mode only, with the humidity set to maximum, and followed a deep cleaning path with lines spaced 15 cm apart instead of 20 cm. First pass: the remaining water was measured by manual sponging and weighed, revealing 2 liters absorbed (40%). Visual observation confirmed the disappearance of the puddles, replaced by a uniform wet film. Second pass: an additional 1.8 liters absorbed (76% total). Third pass: 0.9 liters absorbed (94% total). The remaining liter evaporated naturally in 15 minutes. A respectable performance, but one that could be improved, demonstrating the physical limitations of a robot when faced with a massive flood.
Testing the orange mop on stubborn dirt: 100 grams of melted chocolate was spread over 1 square meter of tiled floor and left to dry for 4 hours to form an adherent crust. The robot was run in intensive mode with maximum humidity. Timing and observation: the first pass lasted 3 minutes with a clearly audible scraping sound. Weighing the remaining crust: 30 grams remained (70% removed). Second pass: 10 grams remained (90% removed). Third pass: estimated residual traces of 2 grams (98% removed). For comparison, we repeated this test with a standard mop from the X50 Ultra Complete: 40 grams remained after three passes. The orange scrubbing mop therefore demonstrates 133% greater effectiveness on this type of dirt, validating the benefits of its specialized design.
Additional test with a ketchup stain: 50 grams of ketchup were spread on tiles and left to dry for 3 hours. Three passes with an orange mop resulted in a residual trace of 5 grams (90% removed). The same test with a standard mop yielded 15 grams of residual trace (70% removed). Performance difference: 29% in favor of the orange mop. While the advantage remains significant on this type of greasy stain, it is not as dramatic as on dried chocolate.
The pressure exerted by the rotating mops was empirically evaluated. An attempt to slide a credit card (0.76 mm thick) under the rotating mops proved impossible without significant force. An attempt with a sheet of paper (0.1 mm thick) was possible with firm traction, resulting in a crumpled sheet. This simple experiment demonstrates high contact pressure, estimated at several hundred grams per mop. The rotation speed, measured using a high-speed camera, was 172 revolutions per minute, slightly higher than the advertised 165 revolutions per minute.
One crucial point deserves emphasis: the phase-change material integrated into the blue and orange mops, designed to maintain a temperature of 40°C during cleaning, fails to deliver on its promise. Measurements taken with a thermal imaging camera immediately after leaving the station showed 45°C on the half of the mop in contact with the heating element, and 26°C on the other half. A further measurement after three minutes of floor washing revealed 19°C, the ambient temperature of the tiles. The thermal material, despite its presence (the blue mop exhibits significantly greater rigidity than the gray one, indicating a multi-layered structure), does not retain heat for more than a few dozen seconds. This technical failure is a disappointment, especially since the manufacturer's marketing materials heavily emphasize this point.
Self-maintenance system: worrying gaps between promises and reality
The docking station centralizes four automatic functions: emptying the dustbin, washing the mop heads, drying the mop heads, and storing/changing the three pairs of pads. We tested each function with particular rigor, especially the washing and drying temperatures, which are advertised at 100°C.
The mop washing process follows a standardized cycle: water is pumped from the clean water tank, heated by two heating elements integrated into the base, the mops are pressed and mechanically rubbed against this heated base for 3 to 7 minutes depending on the selected setting (quick, standard, or deep), and then the dirty water is drawn into the dedicated tank. To measure the actual temperature, we equipped the station with a Flir thermal imaging camera (a model used for professional thermal diagnostics, with an accuracy of ±2°C) and took readings at regular intervals during the washing cycle.
Measurement results: Initial temperature of the pumped cold water: 18°C. Temperature after 2 minutes of heating: 38°C. Maximum temperature reached after 4 minutes: 45°C. Average temperature during the mechanical scrubbing phase: 42°C. These values, measured five times over different cycles, were consistently identical to within ±2°C. None of the measurements approached the 100°C claimed by the manufacturer. The difference between the marketing claim and the measurable reality is 55°C, representing a 55% discrepancy. This significant discrepancy constitutes a serious breach of the transparency that consumers have a right to expect.
From a strictly sanitary standpoint, a temperature of 60°C represents the minimum pasteurization threshold required to eliminate the majority of common pathogenic bacteria. At 45°C, washing adequately removes visible dirt but does not reach this level of disinfection. For households housing immunocompromised individuals or in situations requiring enhanced hygiene (contamination by animal matter), this insufficient temperature can pose a problem. However, it should be noted that during our three-week test, no unpleasant odors appeared despite situations of extreme dirtiness, suggesting that 45°C is sufficient for standard household cleaning.
The mop heads are dried by circulating hot air in the storage compartment for a configurable duration of 3, 4, or 5 hours. Thermal measurements during the drying cycle: temperature at the bottom of the station, where the hot air is blown: 52°C on average (range 48-54°C). Temperature at the level of the mop heads themselves: 38°C on average (range 35-42°C). Temperature of the air expelled from the compartment: 28°C. These measurements, repeated over ten drying cycles, proved remarkably stable. Here again, the advertised 100°C is pure marketing fiction. The difference is at least 48°C, representing a 48% discrepancy with the stated specifications.
Additional observation: the heating elements integrated into the base, visible when the door is opened and easily identifiable by their copper color, are not used during the drying process. Direct thermal measurement of these elements during a drying cycle: 22°C, which is the ambient room temperature. Drying is therefore carried out solely by hot air generated by a separate system (probably an electric heating element coupled with a fan located at the rear of the unit), a less efficient system than the direct contact heating that could have been implemented using the heating elements in the base.
Despite these moderate temperatures, the drying process completed correctly within the stated timeframe. Complete drying test: after intensive washing of the mops (which were very dirty following our chocolate test), a 3-hour drying cycle was programmed. Verification at the end of the cycle, with absorbent paper pressed against each mop: the blue and orange mops were perfectly dry (no detectable moisture), while the gray mop showed a slight residual moisture, estimated at 5% (complete drying achieved after 4 hours). These results, reproducible across all our tests, validate the drying efficiency despite a temperature lower than advertised. The system therefore fulfills its function, but the manufacturer's communication misleads consumers regarding the technical means employed.
The automatic emptying of the dustbin works on the classic principle: the robot vacuums the contents of its bin into the 3.2-liter bag located in the charging station. This operation takes approximately 15 seconds and the noise level is measured at 78 decibels at a distance of 1 meter. We removed the bag after two weeks of intensive use: it contained about 150 grams of various debris and was two-thirds full. No significant residue remained in the robot's bin, confirming the efficiency of the suction. The debris compression system, which Dreame mentions as densifying the bag's contents, appears to be working effectively: the volume occupied by 150 grams of debris was less than what would be expected without compression.
The self-cleaning system of the station's base was observed daily. During our three-week test, which included situations involving significant soiling (chocolate, ketchup, reconstituted mud), the base remained clean without any manual intervention on our part. The large-diameter drain pipe effectively directs all residues straight into the dirty water tank without obstruction. Only a slight limescale film began to appear after two weeks, a normal phenomenon related to the hardness of the mains water (our test water had a hardness of 25°TH). Descaling is therefore necessary monthly or bi-monthly, depending on the local water hardness, an operation made easier by the base's reasonable accessibility despite its non-removable design.
Power consumption was measured using a recording wattmeter during a complete cycle (cleaning 120 m², washing the mops, and drying for 4 hours). Total consumption measured: 1.18 kWh. Breakdown: 0.12 kWh for cleaning (robot + standby station), 0.21 kWh for washing the mops (peak at 650 watts for 20 minutes), and 0.85 kWh for drying (70 watts for 4 hours). Based on daily use, annual consumption is 431 kWh, or €73 at the regulated EDF tariff of €0.17 per kWh. In standby mode, the station consumes 2.1 watts, or 18 kWh per year (€3). These values are within the segment standard, slightly lower than the Roborock S8 MaxV Ultra (460 kWh per year) and higher than the Ecovacs X2 Omni (395 kWh per year).
User experience: advanced automation but learning required
Initial setup requires downloading the Dreame Home app, available on iOS and Android. Pairing is done by scanning a QR code on the robot, an operation completed in 8 seconds during our test. The app then prompts you to run an initial map, an essential step before any use. This map, automatically created by the robot as it scans all accessible rooms, took 11 minutes for our 120 m² area comprising eight rooms.
The generated map proved remarkably accurate. Digital overlay of the robot's map onto the house's architectural plan revealed an average discrepancy of 3 cm in the dimensions of each room, an excellent precision attributable to LiDAR technology. Automatic zone segmentation correctly identified seven out of eight rooms. The only error: a portion of the hallway (2 m²) was included in the adjacent bedroom. This was corrected in 30 seconds using the application's merge/separate tool. Automatic room type recognition worked convincingly: kitchen, bathroom, living room, and bedrooms were correctly categorized, allowing the robot to spontaneously suggest the appropriate mop for each type. This artificial intelligence eliminates the need for tedious setup.
The Dreame Home app is structured into thematic sections: suction settings (four power levels), humidity settings (five levels), mop washing frequency (per hour, per room, or per square meter), carpet management (ignore, vacuum while lifting the mops, or return the mops to the station), detergent solutions (activation/deactivation for each compartment with three dosage levels), and automatic mop switching (assignment of type per room). The abundance of available settings demonstrates a desire to offer precise control, but may confuse users unfamiliar with this type of interface.
Clean Genius mode offers a simplified approach for users who dislike manual settings: this automatic mode entrusts the robot with complete cleaning management, dynamically adjusting suction power and humidity based on the detected dirt level. We tested Clean Genius mode for a week, yielding satisfactory results on everyday dirt, although cleaning time was on average 28% longer compared to optimized manual settings (87 minutes vs. 68 minutes for our 120 m² area). This increase is due to the safety margins built into the automated system and the additional passes made in areas detected as dirty.
The scheduling function allows you to set recurring cleaning sessions according to a calendar. We programmed a daily cleaning at 9 a.m. on weekdays with vacuuming and mopping in standard mode, and a cleaning at 11 a.m. on weekends in intensive mode. For two weeks, all cleanings started on time without fail. This reliability in automatic execution is a real convenience for users who work outside their homes.
The automatic mop changing system works reliably. The complete sequence is timed as follows: the robot leaves the area being cleaned, travels to the charging station (distance varies depending on the configuration), drops off the used mop (20 seconds), performs a quick wash of that mop (3 minutes in standard mode, 5 minutes in deep cleaning mode), selects the appropriate new mop (15 seconds), dampens it (30 seconds), and moves on to the next area. Total measured time: 4 minutes 25 seconds in standard mode, 6 minutes 30 seconds in deep cleaning mode. This time represents an additional cost compared to a conventional robot, but remains acceptable given the benefit provided. During our three weeks of testing, no selection errors were observed: the robot consistently chose the appropriate mop according to the configured room type.
One point to note concerns the automatic detection of low-profile areas for Lidar retraction. On our sofa (under-frame height of 10.2 cm), the Lidar retracted correctly on 8 out of 10 passes. On two passes, it remained extended, forcing the robot to backtrack. Manually adjusting the settings in the app to force retraction under this piece of furniture resolved the issue. This imperfect automatic detection therefore requires a learning phase and manual adjustment, an operation that is not intuitive for novice users.
The dashcam provides a live video feed accessible via the app. Measured image quality: estimated resolution of 640×480 pixels, frame rate of 15 frames per second, latency of approximately 2 seconds between actual movement and display. This modest quality is sufficient for identifying a pet or checking progress, but it is not suitable for reading text or identifying fine details. The front LED light activates automatically in dark areas, ensuring constant visibility.
The provided documentation is limited to an 8-page printed manual in French, covering only the basic operations. The complete manual, downloadable via a QR code, totals 47 pages and covers all functions. This digital approach is justified by environmental considerations, but penalizes users who are not comfortable with PDF files on smartphones. Integrating interactive tutorials directly into the application would have facilitated learning.
Navigation and obstacle detection: effective but could be improved on low objects
The Dreame Matrix 10 Ultra's LiDAR navigation system relies on a laser rangefinder that rotates 360° and measures distances to obstacles 2,000 times per second, according to the manufacturer's specifications. This technology, considered the most reliable on the market, is complemented by a front-facing camera and 3D sensors for precise obstacle detection. We subjected this system to a standardized testing protocol designed to evaluate its performance in various configurations.
Mapping test on a complex surface including furniture of varying heights, corners, and narrow areas: mapping time measured at 11 minutes for 120 m². Accuracy of the resulting map verified by overlaying it with the architectural plan: average deviation of 3 cm on linear dimensions, average deviation of 0.8° on angles. This millimeter-level accuracy validates the quality of the Lidar system. The generated map displays a two-dimensional representation with a clear distinction between accessible areas (in white), permanent obstacles (dark gray walls), and detected furniture (in light gray).
The retractable Lidar is a valuable asset for accessing areas with low ceilings. Tested under a TV stand 9.8 cm high (precisely measured), the robot automatically detected this low height, retracted its Lidar, reducing its overall height to 8.9 cm, activated its front LED light, and proceeded under the stand. Navigation under this stand was efficient, accurately mapping the accessible area and demonstrating the system's ability to seamlessly switch between Lidar and camera navigation. However, as mentioned earlier, this automatic detection proves imperfect in certain configurations, requiring manual adjustments.
Obstacle detection was performed according to a rigorous protocol. Ten objects of varying types and sizes were placed on the floor in an empty room: two electrical cables (black and white, 5 mm diameter), a rolled-up sock (4 cm high), a shoe (8 cm high), a wooden cube (5 cm sides), a remote control (2 cm high), a phone (0.9 cm high), a smartwatch (1.2 cm high), and six dummy droppings made of modeling clay in different sizes (3 to 8 cm in diameter) and colors (light yellow, dark yellow, brown). The robot was then started in full cleaning mode.
Daytime results (natural light, 800 lux measured): Black and white cables detected and avoided with a 3 cm margin. Sock detected and avoided with a 5 cm margin. Shoe detected and avoided with a 4 cm margin. Cube detected and avoided with a 3 cm margin. Remote control lightly touched with the side brush, moved 2 cm. Phone directly touched by the chassis, moved 5 cm. Watch struck by the chassis, moved 8 cm. Of the six dummy droppings: all six detected and avoided with a minimum 5 cm margin, 100% score.
Nighttime results (complete darkness, 0 lux): cables, sock, shoe, and cube were detected and avoided with the same performance as during the daytime test. The remote control, phone, and watch were touched as during the day. Of the six droppings: five were detected and avoided. The 3 cm diameter dropping (shown in light yellow), placed in a corner, was not detected, as the robot lightly drove over it. Score: 83%. These results validate the effectiveness of the detection system on objects taller than 4 cm, but reveal its limitations with very low objects. This physical limitation is explained by the position of the sensors approximately 10 cm from the floor, an insufficient height to detect objects between 1 and 3 cm. This shortcoming, common to almost all current robot vacuums, is worth noting because it can cause inconveniences: remote controls pushed under furniture, phones bumped, and watches moved.
Obstacle clearance was tested with blocks of varying heights: 2 cm (cleared without difficulty), 3 cm (cleared after a slight hesitation), 4 cm (cleared in three out of five attempts, failed after several unsuccessful attempts), and 5 cm (systematic failure after three attempts). These results generally correspond to the stated specifications (4.5 cm maximum), with a margin of tolerance explained by the nature of the obstacle (material, angle of attack), which influences clearance. The optimized obstacle clearance mode, which can be activated in the application, improves the success rate on 4 cm obstacles by first lifting the chassis and then moving one wheel at a time. With this mode activated, the 4 cm block was successfully cleared in 100% of attempts.
The extendable side brush was observed during several cleaning sessions. The articulated arm deployed approximately every 20 seconds when cleaning along baseboards, systematically when passing under low furniture or doors, and occasionally when corners were detected. The maximum measured extension distance was 4.2 cm from the robot's chassis. Corner cleaning effectiveness was verified after two weeks of use without manual intervention: no visible dust accumulation in the corners of the house, confirming the system's effectiveness.
Cost of use and pricing strategy: an investment to be put into perspective
The Dreame Matrix 10 Ultra is priced at €1099, placing it firmly in the ultra-premium segment. To assess the value of this investment, we calculated the total cost of ownership over five years, the estimated lifespan for this type of device based on the reliability data of previous generations.
The necessary consumables include: mop heads (machine washable but should be replaced approximately every 6 months depending on usage), HEPA filters (to be replaced every 4 months), side brushes (to be replaced every 6 months), and dust bags (duration varies depending on the household, estimated at 8 weeks for a household without pets, 4 weeks with pets). Prices observed on the official Dreame website in November 2024 were as follows: a set of six mop heads (two of each type) for €59.99, a set of four HEPA filters for €19.99, a set of four side brushes for €14.99, and a set of twelve dust bags for €29.99.
Calculating the annual cost of consumables for a pet-free household, with monthly machine washing of the mop pads to double their lifespan: mop pads (one set per year) €60, filters (three sets per year) €60, side brushes (two sets per year) €30, dust bags (four sets per year) €120. Total: €270 per year. For a pet-friendly household requiring more frequent changes: €420 per year. These amounts are in the upper range for this segment, slightly higher than the Roborock S8 MaxV Ultra (€240 per year without pets) and lower than the Ecovacs X2 Omni (€310 per year without pets).
Detergent solutions represent an additional expense. Consumption measured over three weeks with standard dosage and daily use: 180 ml of standard solution, 95 ml of odor-eliminating solution, 45 ml of parquet solution. Annual extrapolation: 3.1 liters of standard solution, 1.6 liters of odor-eliminating solution, 0.8 liters of parquet solution. Cost on the official website: standard solution €19.99 per liter, odor-eliminating solution €24.99 per liter, parquet solution €24.99 per liter. Total annual cost: €122. This expense can be eliminated by using only plain water, with an estimated 15% loss of effectiveness on greasy stains according to our comparative tests.
Electricity consumption, measured at 431 kWh per year for daily use, represents a cost of 73 euros at the regulated tariff of 0.17 euros per kWh. This amount varies depending on the applicable tariff (peak/off-peak hours tariff, alternative contract) and the actual frequency of use.
Total cost over five years: purchase price €1099 + consumables (5 years) €1350 (270 x 5) + detergent solutions (5 years) €610 (122 x 5) + electricity (5 years) €365 (73 x 5) = €3424. That's €685 per year, or €1.88 per day. This calculation does not take into account a potential breakdown requiring out-of-warranty repair, an event whose probability remains difficult to predict given the limited experience with this specific model.
Comparison with the cost of a cleaning lady: a weekly 2-hour service at an average rate of €15 per hour in the Paris region = €1,560 per year, or €7,800 over five years. The savings achieved with the robot therefore amount to €4,376 over five years, or €875 per year. This substantial saving financially justifies the investment for households currently using a cleaning service. However, a robot does not completely replace human intervention for tasks not directly on the floor (dusting furniture, cleaning windows, maintaining bathrooms), a point to keep in mind when considering the cost.
Competitive positioning: The Roborock S8 MaxV Ultra (€1099) offers 10,000 Pascals of suction power (three times less), 4- and 3.5-liter tanks (25% smaller), but includes an articulated mop arm for cleaning right up to baseboards, a feature absent from the Dreame. The Ecovacs Deebot X2 Omni (€1199, €100 more) boasts 8,000 Pascals of suction power, 4- and 3-liter tanks, but features a square design for easier corner cleaning. The Dreame X50 Ultra Complete (€799, €300 less) offers performance very close to the Matrix 10 Ultra, except for the interchangeable mop system, which is absent on this model. The central question then becomes: does the interchangeable mop system justify an extra €300?
Our answer, based on our testing: for homes with a wide variety of floor types and soiling patterns (kitchens with frequent grease spills, bathrooms prone to humidity, pets requiring odor control, hardwood floors needing special care), the system provides measurable added value that justifies the extra cost. The orange mop's 133% superior effectiveness on dried chocolate and the gray mop's 30% greater effectiveness in absorbing water are tangible advantages. For other user profiles (apartments with uniform flooring, minimal specific soiling, no pets), the X50 Ultra Complete offers superior value for money.
Dreame Matrix 10 Ultra
✔ Validated suction power: 30,000 Pascals measured, 100% of 200g of debris vacuumed in 3 passes
✔ Specialized and effective mops: Orange scrubber 133% faster, grey 30% more absorbent
✔ Giant tanks: 5.5L clean water + 4L dirty water = 200m² autonomy without intervention
✔ Anti-tangle brush: No tangled hair after 3 weeks of intensive testing
✔ Self-cleaning station: Base remained clean without manual intervention throughout the test
✖ Temperatures lower than advertised: Wash at 45°C instead of the advertised 100°C (-55%)
✖ Inefficient thermal material: Mops cooled to 19°C in 3 minutes, not the promised 40°C
✖ Low obstacle detection faulty: Remote controls and phones are systematically pushed aside
✖ Heavy tanks: 6 kg full of clean water, difficult for elderly people
✖ High noise level: 72 dB in max mode, above the competition
FAQ
Our tests on varnished parquet flooring over two weeks revealed no scratches. However, we recommend reserving this mop for resistant surfaces (tiles, stone, porcelain stoneware) and using standard blue mops on parquet floors. The abrasive nylon could damage delicate varnishes over the long term.
Our analysis reveals that the interchangeable mop system offers 133% greater effectiveness on tough dirt and 30% better water absorption. For homes with a busy kitchen and a humid bathroom, this improvement justifies the extra cost. For an apartment with uniform flooring, the X50 Ultra Complete offers better value for money.
During our three-week test, including intensive washing, the mops showed no signs of wear. Extrapolating this data to monthly machine washing at 30°C, we estimate a lifespan of 6 to 8 months per pair for daily use. The orange scrubbing mops will likely reach the lower end of this range. Estimated annual replacement cost: €120.
We timed our interventions over 3 weeks: refilling with fresh water every 3 days (2 min), emptying with dirty water (2 min), weekly sensor cleaning (1 min). Total: 5 min/week. Add monthly dust bag replacement (2 min), machine washing of the mop pads (10 min), and bi-weekly descaling (5 min). Actual total: 8 min/week for all tasks combined.
From a microbiological perspective, 60°C represents the pasteurization threshold. At 45°C, washing cleans adequately but does not reach this level of disinfection. During our three weeks of testing under extreme conditions, no odors appeared, suggesting that 45°C is sufficient for standard domestic use. For enhanced hygiene (immunocompromised individuals, animal organic matter), monthly hand washing in a washing machine at 60°C is recommended.
Our verdict: an excellent robot vacuum cleaner marred by misleading marketing
After three weeks of real-world testing, the Dreame Matrix 10 Ultra delivers mixed results. On the one hand, its suction power more than lives up to the advertised claims: the promised 30,000 Pascals translate into measurable effectiveness on all types of debris and flooring. The interchangeable mop system, far from being a mere gimmick, provides quantifiable added value in the situations for which it was designed: the orange scrubbing mop effectively removes stubborn dirt 133% faster than standard mops, while the gray mop absorbs 30% more water. The large 5.5-liter and 4-liter tanks offer exceptional autonomy, reducing the need for manual intervention. The self-maintenance station performs its functions correctly, keeping the robot operational without requiring daily user intervention.
On the other hand, the discrepancies between marketing promises and measurable realities cast a significant shadow on this otherwise positive picture. The washing and drying temperatures, advertised at 100°C with strong health claims, are capped at 45°C and 52°C respectively. These discrepancies of 55% and 48% are not mere inaccuracies but rather serious breaches of the transparency that consumers have a right to expect, especially for a product priced at €1099. The thermal material meant to maintain the mop heads at 40°C during cleaning fails to perform its function, cooling down to room temperature in three minutes. Dreame should reconsider its communication strategy, especially since the appliance's actual performance is more than sufficient to convince consumers without resorting to questionable exaggerations.
Priced at €1099, this model competes with the Roborock S8 MaxV Ultra at the same price, which, while offering three times less suction power, includes an articulated mop arm for cleaning right up to baseboards. Compared to the Dreame X50 Ultra Complete at €799, which offers nearly identical performance without the interchangeable mop system, the choice will depend on the layout of your home and the type of dirt you encounter. Our analysis of the total cost of ownership over five years reveals an investment of €3424, including consumables and electricity, or €685 per year—four times less than the cost of a weekly cleaning service.
Our recommendation is primarily aimed at large households (over 100 m²) with varied floor types and levels of soiling: kitchens with frequent grease spills requiring the orange mop, humid bathrooms needing the absorbent gray mop, homes with pets benefiting from odor-control solutions, and hardwood floors requiring specific maintenance. In this context, the Matrix 10 Ultra offers advanced automation and measurable performance that justifies its premium price. On the other hand, single people or couples without pets living in an apartment with uniform flooring would be better off with the X50 Ultra Complete, which is €300 cheaper and offers 90% of the performance without the sophisticated interchangeable mop system.
Ultimately, the Dreame Matrix 10 Ultra is an excellent robot vacuum and mop whose suction and mopping performance surpasses most of what's currently on the market. Its interchangeable mop system is a relevant innovation that addresses real needs. However, the manufacturer's communication, riddled with exaggerations about temperatures and the thermal material, betrays a questionable marketing strategy that undermines the overall credibility. At this price point, consumers deserve complete transparency. Dreame has an excellent product; it just needs to adopt a communication strategy that matches its technical achievements.
"To learn more about the best models, see our complete comparison of robot vacuum cleaners ."
Our best robot vacuum cleaner tests and comparisons
The Dyson Vis Nav 360 impresses with its record-breaking 65W suction power, comparable to
OVERALL SCORE: 16/20 The Ecovacs Deebot X2 Omni represents a bold break in the world of
VERDICT IN BRIEF OVERALL SCORE: 15.5/20 Roborock strengthens its offering in home robotics with
Dreame Technology, a Xiaomi subsidiary since 2017, is accelerating its offensive on the French market.
"Robot vacuum cleaners are revolutionizing floor cleaning. They promise powerful suction and efficient washing."
When I took the MOVA Z50 Ultra out of its box, I didn't expect