Which cars are in the best condition for their age and mileage?

Pass rates are the usual way to judge how well a car lasts, and they are misleading. A three-year-old car passes its MOT far more often than a twelve-year-old one, whoever built it, and a car that covers 2,000 miles a year is in better condition than one doing 15,000. Rank models on pass rates and you mostly rank them on age and use.

So we built a measure that takes both out. The UK Motoring Reliability Score compares each model with what cars of the same age and annual mileage actually record, from 24.1 million MOT tests in 2025 where the mileage could be measured. 50 is exactly as expected. Higher is better. The full method is published, and every model’s score is on its own page.

The case for it, in two cars

A Dacia Sandero passes its MOT far more often than a Citroën DS3, and is in worse condition than it should be, while the DS3 is closer to par.

	Dacia Sandero	Citroën DS3
First-time pass rate	69.1%	51.8%
Average age at test	7.5 years	12.1 years
Failure rate	33.8%	49.0%
Expected for that age and mileage	22.0%	33.3%
Reliability Score	26	29

The Sandero looks the better car on the raw number and is the worse one on the fair comparison. Its cars are young, and young cars are not supposed to fail a third of the time.

That is the whole argument for the score in one row of a table.

Which cars score highest?

The Aston Martin DB9 tops the table on 88, failing 9.0% of its MOTs against 33.9% expected for cars of its age and mileage.

Model	Score	Fails	Expected
Aston Martin DB9	88	9.0%	33.9%
Aston Martin Vantage	87	6.6%	22.2%
Bentley Bentayga	86	5.0%	16.7%
Bentley Continental	85	8.1%	28.4%
Aston Martin DB11	84	5.0%	14.8%
Audi R8	83	9.5%	27.2%
Porsche 911	82	11.0%	30.1%

The honest caveat

That list is expensive cars, and it would be dishonest to present it as proof that expensive cars are built better. A DB9 covers 1,406 miles a year. It is garaged, serviced on time, and driven on dry Sundays. The score standardises for mileage, which removes some of that advantage, but not the standard of maintenance.

What the score measures is the condition of cars on the road. That is the product of how a car was built and how it has been kept, and for a buyer both matter: you are buying the car in front of you, not the factory that made it.

The best and worst cars people actually buy

Restrict it to models with more than 50,000 tests a year, the cars you will meet in the classifieds, and the table becomes useful.

Highest scoring	Score	Lowest scoring	Score
Porsche 911	82	Dacia Sandero	26
Audi Q5	68	Citroën DS3	29
Volkswagen T-Roc	67	Renault Clio	29
Audi Q2	67	Dacia Duster	32
Audi A6	66	Citroën C4	32
BMW 5 Series	65	Vauxhall Mokka	35

The Renault Clio is the one to note: 306,492 tests, more than almost any model on the road, and a score of 29.

By class, which is how people actually shop

Nobody cross-shops a DB9 and a Sandero. Within a class, the spread is still wide.

Class	Highest scoring	Score	Lowest scoring	Score
Supermini	MINI Cooper	60	Dacia Sandero	26
Small family	Toyota Prius	64	Citroën C4	32
Small SUV	Volkswagen T-Roc	67	Vauxhall Mokka	35
Family SUV	Honda CR-V	62	Peugeot 3008	38
Executive	BMW 5 Series	65	Tesla Model 3	38
City car	Toyota Aygo	51	Fiat 500	37

Related: every model ranked by score

The Prius proves the point about mileage

The Toyota Prius scores 64 while covering 13,326 miles a year, more than twice the average car.

Because it does the mileage of a taxi, 34.5% of its tests would be expected to fail. It fails 26.2%. On a raw pass rate the Prius looks ordinary; standardised for the work it does, it is one of the best cars in its class.

The Porsche 911 is the mirror image. Its 2,149 miles a year make its 89.1% pass rate much less impressive than it first looks, which is why its score, at 82, sits below the DB9’s despite a better raw pass rate than most of the table.

Electric cars are not the story people expect

The Renault Zoe scores 20, the second lowest of any model. It fails 24.6% of its tests against 15.4% expected for cars that young.

The reason is not the battery or the motor, neither of which the MOT tests. Tyres account for 40% of Zoe failures. Electric cars are heavy and deliver full torque from a standstill, and tyres pay for it. The Tesla Model 3 tells a milder version of the same story: 16.6% of tests fail against 15.0% expected, and a score of 38.

This is a real finding, not a knock on electric cars: it says that the running costs of an EV include tyres more often than an owner expects, and that a used one deserves a close look at all four corners.

What the score does not do

It does not measure breakdowns, warranty claims, or what a repair costs. A model that fails often on light bulbs is not worse to own than one that fails rarely on gearboxes, and the MOT cannot tell the difference. Cars under three years old are not tested at all, so nothing here covers a nearly-new car.

And it is a measure of the fleet, not of your car. A well-kept Sandero will sail through; a neglected 911 will not. The score tells you the odds, not the outcome.

How it is worked out

Each model’s failure rate and dangerous-defect rate are compared with the national rate for cars in the same seven age bands and five annual mileage bands, weighted by where that model’s own tests fall. Failure rate counts for three quarters of the score and dangerous defects for the remaining quarter. The published figure comes with a 95% range, and no score is published where that range is wider than five points either side.

Annual mileage comes from consecutive MOT odometer readings 300 to 430 days apart. That is why 102 models with published MOT figures have no score: too few of their tests could be paired to measure it.

The full method sets out every threshold and every reason a score is withheld.