Qrevo Edge 2 vs Saros 10R vs Saros 20: The Real Pick

Match the decision to what actually breaks down on real floors, not to which product page has the more convincing FlexiArm copy. In independent lab testing, obstacle avoidance separates these three robots far more than edge-hugging does: the Saros 10R cleared a perfect 24 of 24 obstacles in VacuumWars testing, the Saros 20 cleared 17 of 24, and the Qrevo Edge 2 cleared only 14 of 24 — below both the category average and its own predecessor’s score. If cords, toys, or pet mess are a fact of life on your floors, that gap matters more than anything FlexiArm does.

Skip to the picks

Why FlexiArm feedback looks contradictory

“FlexiArm” isn’t one mechanism — it’s a marketing name Roborock has attached to at least three different physical designs. The Qrevo Edge 2 uses a FlexiArm Arc Side Brush that swings a curved brush and one spinning mop pad outward. The Saros 10R uses a FlexiArm Riser Side Brush with a different extension geometry. The Saros 20 pairs its arm with a dock that swaps the mop pads themselves mid-clean. Reviewers testing each generation are describing different hardware, which is why search results for “FlexiArm” contain both praise and complaints depending on which model and which model-year is being discussed.

Some of the confusion also traces back to the original Qrevo Edge — the predecessor to the Edge 2. TechRadar’s review of that model found it “stumbles in the one area it should shine,” reporting a roughly half-inch gap between where the mop reached and where the baseboard began. That review still ranks for Qrevo Edge searches and gets read by people comparing the Edge 2, even though the two are different products with different FlexiArm hardware.

The upshot: FlexiArm’s reach isn’t really the variable worth worrying about across these three. All three get a side brush and mop pad close to a wall. What varies — and what independent testing actually measured — is what happens when the robot has to get past something to reach that wall in the first place.

The numbers that separate these three

Roborock also publishes an “objects recognized” figure for each model’s AI camera — 280+ for the Qrevo Edge 2, 201 for the Saros 20, and 108 for the Saros 10R. That catalog size does not predict the lab result: the Edge 2 recognizes the most object categories on paper and scored the worst in actual obstacle-avoidance testing, while the 10R recognizes the fewest categories and scored a perfect run. If you’re using “number of objects recognized” as a shorthand for how well a robot avoids your living room floor, VacuumWars’ own test data says that shorthand is unreliable.

Cords, toys, and pet mess push the decision toward the Saros 10R

For a floor that regularly has charging cables, kids’ toys, or shed pet hair on it, obstacle avoidance is the spec that determines how much cleanup you do after each run, not suction. The Saros 10R’s perfect 24/24 score wasn’t a fluke of one test — VacuumWars ran it through five separate obstacle scenarios (pet waste, toys, cloth, cords, and a combined “torture test”) and it passed every one, at a time when most tested robots failed at least one. TechGearLab’s independent lab separately measured its edge cleaning by scattering coffee grounds against a wall and found it picked up debris to within 0.1 inches of the baseboard, calling the FlexiArm Riser Side Brush’s edge coverage claim accurate in practice.

The tradeoff on the 10R is threshold crossing and, per aggregated owner reports collected by VacuumWars, navigation quirks that show up more in everyday homes than in a controlled lab: roughly 31% of reported issues involve mapping behavior like re-scanning walls or wandering, and the robot’s real-world threshold performance undershoots its own rating — it cleared 26 mm boards in testing but couldn’t clear a single-tier board around 36 mm, despite a marketed 40 mm figure. If your home is a single level with no raised thresholds between rooms, that gap won’t matter. If you have a saddle strip between hardwood and tile, it will.

Uneven thresholds and thick rug push the decision toward the Saros 20

The Saros 20’s AdaptiLift Chassis 3.0 is the only one of the three built around a leg-and-wheel mechanism designed specifically for thresholds and pile carpet, rather than a flush, ultra-slim body. In testing, it crossed a 57 mm threshold — more than double what the Saros 10R managed and well beyond the Qrevo Edge 2’s rated 40 mm double-layer figure. Roborock rates the chassis for double-layer thresholds up to 87.9 mm (a 45 mm step plus a 43 mm step) and for dynamic cleaning on carpet pile up to 3 cm. It also carries the highest rated suction of the three at 36,000 Pa, which showed up in VacuumWars’ carpet deep-clean test as 89% sand removal against a 78% category average.

The cost of that mechanical complexity is a slightly slower clean and a mid-pack obstacle-avoidance score: navigation efficiency measured 0.67 square meters per minute against a 0.70 category average — the only one of the three to test below its own category average on that metric — and its obstacle-avoidance run landed at 17 of 24, ahead of the Edge 2 but well behind the 10R’s perfect score. It’s also the newest and mechanically busiest of the three, with legs and wheels doing work that a flush-bodied robot doesn’t need to do at all, so it has the shortest ownership track record of the three models discussed here.

Flat, consistently clear hard floors are the Qrevo Edge 2’s actual case

Strip away the obstacle-avoidance and threshold numbers and the Qrevo Edge 2 has the best raw cleaning scores of the three. VacuumWars measured a 94% carpet deep-clean score (the second-highest the outlet had recorded at review time), 89.5% pet hair pickup, and a combined mopping score of 40.8 against a 20.3 category average — roughly double. Its navigation was also the most efficient of the three tested, covering 0.81 square meters per minute against a 0.71 average, finishing runs about 16 minutes faster than average. Its RetractSense LiDAR turret retracts automatically for low-clearance furniture while keeping full mapping precision everywhere else, and at 3.14 inches tall it slides under the same furniture the Saros 10R does.

None of that changes the obstacle-avoidance number. A 14/24 score, below both the category average and the original Qrevo Edge’s own 16/24, means this is the model most likely to still run into something on a floor that isn’t kept clear. It’s the right pick for a home with flat floors, no high thresholds, and floors that are actually vacuumed clear of stray cords and toys before a cleaning cycle — the scenario where its cleaning and mopping numbers get to matter instead of getting undercut by a collision.

What the arm itself delivers on each model

Since the arm mechanism is the detail that started this comparison, it’s worth being direct: none of the three has a documented FlexiArm reliability problem in independent testing. The complaint about visible gaps along baseboards belongs to the original Qrevo Edge, not the Edge 2. The Saros 20’s FlexiArm implementation was called the best edge cleaning Android Authority had tested on a robot vacuum, getting the sweeping brush and mop pads right to the wall and around furniture legs. The Saros 10R’s version was independently verified down to a tenth of an inch. The Edge 2’s redesigned Arc Side Brush is what VacuumWars credited for its strong mopping and carpet scores in the first place.

If FlexiArm reach was the thing giving you pause, it shouldn’t be. The number worth checking before you buy any of these three is the obstacle-avoidance and threshold data above, not the side-brush marketing.

The dock and mop differences worth knowing

The docks scale roughly with price tier. The Saros 20’s RockDock washes mop pads at up to 212°F (100°C) and swaps a used pair of mop pads for a clean pair mid-clean without any manual step. The Saros 10R’s 10-in-1 dock washes at 80°C with 131°F hot-air drying and handles up to 60 days of dust collection automatically. The Qrevo Edge 2’s dock washes at 176°F (80°C), dries at 131°F (55°C), and holds roughly 65 days of debris in its disposable bag. All three auto-empty and auto-refill detergent or clean water; none require you to touch a dirty mop pad by hand under normal operation.

If you’re weighing an Edge-series robot specifically, dock reliability is worth a separate look before you buy — see Qrevo Edge Dock Failure: Fix It or Return It for what tends to go wrong with the auto-empty and wash cycle on this product line and when it’s worth an RMA versus a fix.

Before you buy either Saros model

Two practical details apply to the Saros line and not the Qrevo Edge 2. First, the Saros 10R only supports 2.4GHz Wi-Fi — if your router broadcasts a combined SSID or 5GHz-only guest network, you’ll need to set up a dedicated 2.4GHz network before setup will complete. Second, both Saros models rely on their respective AdaptiLift mechanisms for threshold work, and Roborock has iterated that mechanism across generations — the Saros 20’s “Chassis 3.0” is a materially different, and per VacuumWars’ measurements more capable, implementation than what shipped on the 10R. Don’t assume threshold performance carries over between the two just because both use the AdaptiLift name.

If your floors are mostly clear and flat, the Qrevo Edge 2’s cleaning and mopping scores make it the one to buy on cleaning performance alone. If cords, toys, or pets are a daily reality, the Saros 10R’s tested obstacle avoidance is the reason to pick it over either alternative. If your home has raised thresholds or thick rug pile the others can’t cross, the Saros 20 is the only one of the three built to handle it, and its suction and threshold numbers back that up in testing.