CarGearBench

We don’t test the gear. We read the people who do — and we show our working.

Every figure names its lab and its conditions · We earn a commission if you buy through our links

Car gear · measured

The number, and where it came from.

A spec sheet tells you a dash cam shoots 4K. It doesn’t tell you whether you can read the number plate of the car in front at night. We collect the measurements that answer that, name who took them, and state what they were taken under.

A dark suburban road at night seen from a windscreen, with the number plate of the car ahead blown out by headlight glare into an unreadable bright rectangle.
Illustration, not evidence — the failure this site measures. A number plate is retroreflective: it throws your headlights straight back into the lens, and past about thirty feet after dark every camera tested loses the characters to the glare.

Nine dash cams. One number plate. After dark.

The best managed thirty feet. Four of the nine read nothing at all — and one of those four records in 4K, while two cameras that did read the plate record in 1440p.

Night range — distance at which the plate stopped being readable
  • Rove R2-4K Dual Pro 30 ft
  • Garmin X310 25 ft
  • Garmin Dash Cam Live 20 ft
  • Garmin X210 20 ft
  • Nextbase 622GW 20 ft
  • Cobra SC 250R no reading
  • Cobra SC 220C no reading
  • Nextbase Piqo 2K no reading
  • Nextbase 322GW no reading

Scale runs to 30 ft, the highest measured value here.

use it MotorTrend, Best Dash Cams for 2026: 9 Tested Picks Ranked by Real-World Performance, 3 April 2026. Cones set at 10 ft intervals ahead of a 2026 Subaru Outback, a second car parked at each interval, and the distance recorded at which its number plate became indecipherable. Run once in daylight and once at night, nine cameras, same car, same plate. Setup time was taken on a stopwatch from opening the box to a mounted, updated, correctly aimed camera.

How it was measured, and which one to buy

What the marks on every figure mean

  • use it Someone independent measured it, we name them, and we state the conditions.
  • marked The number came off the box. We say so every time — a claim is not a measurement.
  • not merged Two labs disagreed. Both are shown. Averaging invents a precision neither has.

The full method