Last updated 1 November 2026.
Key findings
- Ten years on, similar cars diverge sharply. Among cars first used in 2015, 36.0% of Vauxhall Corsas and 34.5% of Ford Fiestas failed first time, against 17.9% of VW Golfs.
- Failure climbs with age. 11.7% of cars failed first time at the first MOT, 22.3% at 7 to 9 years, 31.3% at 10 to 12 years and 42.4% at 16 to 18 years.
- First MOT by make. Of makes with at least 3,000 tests, Citroen (16.6%) and Tesla (15.8%) failed most often at three years old. Honda (6.0%) and Lexus (4.9%) failed least.
- First MOT by fuel. Diesels failed 17.2% of the time, electric cars 13.4%, petrol cars 10.2% and hybrids 9.7%.
- A lower mileage reading is rare. 27,750 of 19.5 million cars (0.14%, about 1 in 700) had a reading at least 1,000 miles lower than at their test two years earlier.
"Initial failure" counts any car that had at least one failing item when first tested, including faults fixed at the test station on the day. We show the final failure rate too, which counts only cars that were not fixed there.
Ten years on: which cars fail most
These are cars first used in 2015 and tested in 2025 at around ten years old. Fixing the registration year compares like with like. Only models with at least 2,500 tests are included.
Highest initial failure rate, 10 models
| Model | Tests | Initial failure |
|---|---|---|
| Citroen Berlingo | 8,362 | 44.9% |
| Citroen DS3 | 5,200 | 44.7% |
| Renault Clio | 13,165 | 44.2% |
| Dacia Sandero | 8,213 | 43.5% |
| Vauxhall Vivaro | 10,532 | 42.5% |
| Renault Megane | 3,609 | 42.1% |
| DS DS3 | 2,520 | 41.9% |
| Renault Trafic | 4,993 | 41.1% |
| Citroen C4 | 14,114 | 38.8% |
| Dacia Duster | 3,261 | 38.8% |
Lowest initial failure rate, 5 models
| Model | Tests | Initial failure |
|---|---|---|
| BMW X5 | 3,328 | 14.8% |
| Audi TT | 4,321 | 15.6% |
| Honda Jazz | 11,251 | 16.2% |
| BMW 520 | 6,526 | 16.7% |
| Honda CR-V | 8,527 | 16.8% |
The range runs from 14.8% for the BMW X5 to 44.9% for the Citroen Berlingo. Three familiar hatchbacks of the period split widely: 36.0% for the Corsa (48,712 tests), 34.5% for the Fiesta (71,326 tests) and 17.9% for the Golf (37,999 tests). DS3 appears twice because DVSA records it under both Citroen and DS.
First MOT: failure rate by make
These are cars first used in 2022 and tested in 2025 at around three years old, for makes with at least 3,000 tests. The tables show the 10 highest and 5 lowest of 30 makes.
Highest initial failure rate, 10 makes
| Make | Tests | Initial failure | Final failure |
|---|---|---|---|
| Citroen | 29,853 | 16.6% | 12.1% |
| Tesla | 38,965 | 15.8% | 14.3% |
| Vauxhall | 70,476 | 14.9% | 10.9% |
| Peugeot | 47,987 | 14.4% | 10.6% |
| MG | 39,135 | 14.4% | 10.7% |
| Ford | 140,347 | 13.6% | 10.0% |
| Renault | 27,241 | 13.1% | 10.3% |
| Dacia | 19,942 | 13.0% | 9.1% |
| Hyundai | 59,131 | 12.9% | 9.0% |
| BMW | 80,399 | 12.6% | 11.0% |
Lowest initial failure rate, 5 makes
| Make | Tests | Initial failure | Final failure |
|---|---|---|---|
| Lexus | 7,701 | 4.9% | 3.2% |
| Honda | 18,875 | 6.0% | 3.5% |
| Suzuki | 13,851 | 7.1% | 4.5% |
| Porsche | 12,059 | 7.4% | 6.2% |
| Mazda | 19,059 | 8.4% | 5.3% |
Tesla's final failure rate (14.3%) is the highest of the 30 makes, because few Teslas that failed were fixed at the test station on the day.
Failure rate by age
All cars aged three or over tested in 2025. The initial failure rate more than triples between the first MOT and the oldest age band.
| Age (years) | Tests | Initial failure | Final failure |
|---|---|---|---|
| 3 (first MOT) | 1,423,242 | 11.7% | 8.9% |
| 4-6 | 5,455,316 | 14.8% | 11.6% |
| 7-9 | 6,939,914 | 22.3% | 18.1% |
| 10-12 | 6,136,638 | 31.3% | 25.5% |
| 13-15 | 4,124,177 | 38.7% | 31.5% |
| 16-18 | 2,766,808 | 42.4% | 34.4% |
| 19+ | 2,458,647 | 39.9% | 32.8% |
First MOT: failure rate by fuel
The same three-year-old cars as above, split by fuel type.
| Fuel | Tests | Initial failure | Final failure |
|---|---|---|---|
| Diesel | 202,453 | 17.2% | 12.8% |
| Electric | 203,728 | 13.4% | 11.0% |
| Other | 24,390 | 12.9% | 11.2% |
| Petrol | 586,020 | 10.2% | 7.9% |
| Hybrid | 295,915 | 9.7% | 7.1% |
"Other" covers fuel types recorded outside the four named ones. We make no claim about why the rates differ. See the caveats below.
How rare is a lower mileage reading?
We compared the odometer readings of 19.5 million cars that had a normal MOT in both 2023 and 2025 (a median gap of 2.0 years). A "lower reading" means the 2025 reading was at least 1,000 miles below the 2023 reading. 27,750 cars, or 0.14%, fell into that group.
The share is higher for a few models, mostly imported Japanese people carriers and some vans. These are the top 10 models with at least 2,000 cars compared:
| Model | Cars compared | Lower readings | Share |
|---|---|---|---|
| Nissan Elgrand | 3,323 | 217 | 6.5% |
| Toyota Alphard | 2,536 | 160 | 6.3% |
| Toyota Estima | 7,361 | 193 | 2.6% |
| Citroen Dispatch | 24,703 | 290 | 1.2% |
| Peugeot Expert | 27,361 | 311 | 1.1% |
| Vauxhall Vivaro 2900 Dynamic S/S | 2,750 | 27 | 1.0% |
| Toyota Prius | 74,471 | 730 | 1.0% |
| Fiat 500X | 27,378 | 249 | 0.9% |
| London Taxis INT TX4 | 4,560 | 39 | 0.9% |
| Peugeot Rifter | 5,058 | 40 | 0.8% |
This table doesn't show that any car's mileage was changed. A reading lower than the earlier test can come from a tester typing error, a missing digit, kilometres entered as miles, a replaced odometer or instrument cluster, or tampering. DVSA notes that kilometre readings were sometimes entered as miles before 2022, and imported cars can still be affected. That may explain the Nissan Elgrand (6.5%), Toyota Alphard (6.3%) and Toyota Estima (2.6%), but the data can't confirm it. Vans such as the Citroen Dispatch (1.2%) and Peugeot Expert (1.1%) are also above average.
Our mileage check guide shows how to read a car's MOT mileage history yourself, and you can check any plate free.
Methodology
- Source. DVSA anonymised MOT tests and results, published on data.gov.uk. We used the 2025 results (all 12 monthly files) and, for the mileage comparison, the 2023 results.
- Population. Test class 4 (cars and light vehicles up to 8 seats, including small vans), normal first-attempt tests only, with a result of Pass, Fail or Pass with rectification at the station. Retests, partial retests, aborted and abandoned tests are excluded. That leaves 29,603,003 tests on 28.4 million vehicles in 2025. Across that population, 27.8% of tests were an initial failure.
- Initial failure rate is Fail plus Pass-with-rectification-at-station, as a share of all tests. Final failure rate is Fail only.
- First MOT. Cars first used in 2022, tested in 2025 at 2.9 to 3.6 years old (1,312,506 tests). Makes need at least 3,000 tests.
- Ten-year-old cars. First used in 2015, tested in 2025 at 9.9 to 10.6 years old (1,419,489 tests). Models need at least 2,500 tests.
- Age bands. Age at test is the test date minus the first use date, for cars aged 2.9 years or more.
- Mileage. For each vehicle, the last 2023 reading above zero is compared with the first 2025 reading above zero. Only cars first used in 2010 or later are included, which avoids five-digit odometer rollover on older cars, and readings must be below 400,000. Models need at least 2,000 cars compared. The 2024 files were not used.
- Confidence intervals. We calculated 95% Wilson confidence intervals for every rate. They are in the CSV files.
- Rounding. Percentages are rounded to one decimal place (two for the overall mileage share).
- Data downloaded and analysed 19 October 2026.
Caveats
- Initial failure doesn't mean unroadworthy. It includes small items, such as a bulb or wiper blade, fixed at the test station.
- We don't know why cars fail. We didn't analyse the reasons for rejection, so we can't say whether a make or model fails on brakes, lights, tyres or anything else.
- Who owns the car matters. Makes and models differ in how many are company cars, taxis, vans or fleet vehicles, and in how they are serviced. The rankings describe the cars on the road in 2025, not how well they were built.
- Model names are the text DVSA records, so trims can split into separate rows and some names merge several generations.
- Mileage readings are entered by testers. Cars not tested in both 2023 and 2025 are excluded, and a transferred registration plate can, rarely, link two different vehicles.
How to use this when buying
- Check the exact car, not the model. Two cars of the same age can have very different MOT records. Read the full MOT history, including past failures and advisories.
- Look at the mileage at every test. A reading that drops, or a long gap, is worth asking the seller about. There is often an innocent reason, so ask before you assume. Our mileage check guide explains what to look for.
- Budget for repairs on older cars. The failure rate rises with age, so a ten-year-old car is more likely to need work at its next test.
- Work through the rest. Our used car checklist covers the paperwork, the test drive and what to ask the seller.
Data and reuse
Journalists and researchers are welcome to use the figures on this page with a link back to it. Email support@carcontrol.uk if you would like the CSV files behind the tables, which include the confidence intervals and further cuts, such as failure rates by model at the first MOT.
Contains public sector information licensed under the Open Government Licence v3.0. Source: Driver and Vehicle Standards Agency (DVSA), anonymised MOT tests and results, 2023 and 2025.