The Headline Number
Every year, thousands of UK learner riders face the same decision: a new Chinese 125 for £1,500-£2,500, or a Japanese alternative for £3,500-£5,000. The forums are full of people saying the Chinese option has improved. But nobody has looked at what actually happens to these bikes at MOT time - at scale, across hundreds of thousands of real-world tests.
Using the full DVLA MOT dataset - over 874 million test records covering every MOT carried out in Great Britain - we ran the largest analysis of Chinese 125cc motorcycle reliability ever conducted in the UK.
The headline: Chinese 125cc motorcycles and scooters fail their MOT at a rate of 26.0%, compared to 21.0% for equivalent Japanese machines (Honda, Yamaha, Kawasaki, Suzuki - filtered to sub-500cc to compare like for like).
That's a real gap. But the headline number doesn't tell the interesting story.
Early-Life Faults, Not a Longevity Problem
Here is the MOT fail rate broken down by years since first registration:
Look at the chart carefully.
Chinese bikes start at 25% and go essentially nowhere. Year 1: 25.0%. Year 12: 23.7%. The line is flat. A Chinese 125 is almost exactly as likely to fail its MOT in its twelfth year as in its first.
Japanese bikes start at 14.3% and climb steadily to 21.5% by year twelve. They age into their problems the way you'd expect any mechanical device to - gradually and predictably.
At year one to three, a Chinese 125 is 61% more likely to fail its first MOT than a comparable Japanese bike (25.8% vs 16.0%). By year twelve, the gap has narrowed to just 10% (23.7% vs 21.5%).
"The curve is telling a story of early-life faults and non-age-related failure patterns - not a gradual wear-and-tear decline."
The Chinese-segment line stays high and flat. That is hard to square with a story where problems mostly accumulate with mileage and age.
By contrast, the Japanese curve rises over time in the way you would expect when wear, use, and maintenance history start to show up in the test record. For the Chinese-segment cohort, that wear-driven pattern is not what dominates the headline fail rate.
What the data shows is a high rate of failure from the first years onward without a matching climb - consistent with issues that show up early rather than steadily worsening with age.
Survivorship bias caveat: some of the narrowing gap at years 10-12 may reflect the worst Chinese bikes leaving the fleet - bikes that fail catastrophically are scrapped and drop out of the data, leaving a self-selected survivor population. This would mean the year-12 Chinese fail rate is slightly flattering. We can't fully control for this with the available data, so the convergence at the tail should be read as directionally correct rather than precise.
What Chinese Bikes Actually Fail On
This is where it gets specific.
MOT defect notes are free text. Testers often describe the same fault in different words - for example "headlamp not working on dipped beam", "nearside headlight inoperative", and "front lamp failure".
We grouped similar descriptions together so counts reflect the underlying fault, not one exact phrase. The technical approach is in our MiniLM embeddings write-up; we also cross-checked against keyword-based counts.
For riders, the takeaway is simple: lighting and lamp-related items show up far more often per test on Chinese 125cc bikes than on the Japanese comparison set.
Chinese 125cc bikes generate 75% more lighting-related defects per MOT test than Japanese equivalents (258.5 vs 147.4 per 1,000 tests, all defect types). Counting only confirmed failures - MAJOR, FAIL, or DANGEROUS - the gap is 92% (166.7 vs 86.7 per 1,000 tests). We quote the more conservative all-defects figure throughout; the narrower MAJOR/FAIL figure is available in the data tables below. Virtually all named lighting defects are MAJOR or outright failures, not advisories.
| Group | All lighting defects / 1,000 tests | MAJOR+FAIL only / 1,000 tests |
|---|---|---|
| Chinese 125cc | 258.5 | 166.7 |
| Japanese ≤500cc | 147.4 | 86.7 |
Advisory-only lighting defects account for under 2% of the named failure types below - this is a real failure pattern, not noise in the advisory column. We verified each defect category against the raw defect_type field in the DVLA data (values: MAJOR, FAIL, DANGEROUS, ADVISORY, MINOR, PRS).
| Defect | MAJOR | FAIL | Advisory % |
|---|---|---|---|
| Headlamp not working on dipped beam | 4,323 | 3,302 | 0.2% |
| Headlamp: does not illuminate on dipped beam | 4,384 | - | 0.0% |
| Rear stop lamp not working | 3,440 | - | 1.4% |
| Registration plate lamp inoperative | 4,926 | - | 0.2% |
| Front stop lamp not working | 2,966 | - | 0.4% |
| Headlamp aim too low | 3,059 | - | 0.2% |
| Headlamp aim too high | 2,233 | - | 0.0% |
Seven of the top thirty individual failure reasons for Chinese bikes are lighting or electrical. For Japanese bikes, lighting appears but doesn't dominate at the same level.
That distribution helps explain the flat age curve on the chart. The pattern points to higher rates of early-life electrical and lighting faults - the sort often associated with connectors, looms, or lamp assemblies - showing up from early tests onward rather than climbing steadily with age.
You do not need a high-mileage history for a headlamp or stop-lamp fault to appear on the record. That sits more comfortably with non-age-related failure modes than with a pure wear story.
The Brakes and Tyres Picture
Chinese bikes also have 77% more DANGEROUS-classified defects as a proportion of all defects than Japanese bikes (3.9% vs 2.2%). The most common dangerous defects are tyre tread below minimum and brake pads worn to critical levels.
| Defect category | Chinese % of defects | Japanese % | Difference |
|---|---|---|---|
| Lighting | 12.2% | 10.6% | +15% |
| Suspension | 11.0% | 9.9% | +11% |
| Brakes | 23.8% | 23.6% | ≈same |
| Drive chain | 7.5% | 9.1% | −18% |
| Tyres | 10.4% | 13.0% | −20% |
| Steering/headbearing | 5.9% | 5.6% | ≈same |
| Exhaust | 5.4% | 4.9% | ≈same |
The counterintuitive result: Chinese bikes have proportionally fewer chain and tyre defects than Japanese. This reflects usage patterns - Chinese 125s are predominantly low-mileage urban commuters. Chains and tyres simply don't see the same wear as on a Japanese bike covering motorway miles. The Chinese bike risks are specific (electrics, safety components), not general.
Not All Chinese Bikes Are Equal (or even from China)
Lumping every budget brand together obscures important differences. The data covers 16 Chinese and Chinese-market brands with meaningful MOT histories:
| Brand | Unique bikes | MOT fail rate ↑ worst first |
|---|---|---|
| JINLUN | 4,979 | 27.0% |
| GENERIC | 4,162 | 24.6% |
| PULSE | 9,937 | 24.3% |
| LIFAN | 6,509 | 23.8% |
| LEXMOTO | 79,629 | 22.3% |
| KEEWAY | 22,737 | 21.4% |
| SINNIS | 12,641 | 20.8% |
| ZONTES | 8,324 | 18.3% |
| SYM | 51,150 | 18.1% |
SYM: The Hidden Outperformer
SYM is Taiwanese, not mainland Chinese - a distinction that matters. The brand has been manufacturing scooters since 1954 and has genuine engineering heritage. We've included them because they compete directly in the same UK market segment at similar prices and are routinely grouped with "Chinese bikes" in buying guides.
SYM's 18.1% overall fail rate sits between Chinese and Japanese. Some SYM models go further:
| Model | Unique bikes | Fail rate |
|---|---|---|
| SYM Joyride | 549 | 16.4% |
| SYM Symply 50 | 230 | 17.0% |
| SYM Megalo | 285 | 17.4% |
| SYM AV | 794 | 17.5% |
The SYM Joyride at 16.4% is more reliable than the Japanese segment average of 21%. This is based on 549 unique bikes across multiple test cycles. At one end of the budget segment: JINLUN failing 27% of the time. At the other: the SYM Joyride beating Honda and Yamaha on reliability. The price difference between them in the UK market is negligible.
Best and Worst Individual Models
For models with at least 200 unique bikes and 500 total tests:
| Make | Model | Unique bikes | Fail rate |
|---|---|---|---|
| SYM | Joyride | 549 | 16.4% |
| SYM | Symply 50 | 230 | 17.0% |
| SYM | Megalo | 285 | 17.4% |
| Zontes | ZT125-G1 | 465 | 17.5% |
| Lexmoto | Aura 125 | 673 | 19.1% |
| SYM | City Hopper | 1,119 | 23.0% |
| Lexmoto | Milano | 1,557 | 22.3% |
| SYM | Jet | 5,884 | 21.7% |
What This Means If You're Buying a 125
If you're buying new: The gap at year one is real and significant. A new Chinese 125 from most brands is 61% more likely to fail its first MOT than a new Japanese equivalent.
In practice, budget for a pre-MOT check and pay extra attention to lights, lamps, and wiring-related advisories and failures - that is where the data says the gap concentrates.
If you're buying used: A ten-year-old Chinese 125 failing at 23.7% is almost statistically indistinguishable from a ten-year-old Japanese bike at 21.5%. The "buy old Japanese because it'll be more reliable" logic holds less weight at higher ages than most people assume.
If you're shopping specifically: The SYM Joyride and Zontes ZT125-G1 are the data-backed standouts. The JINLUN brand has the worst reliability record of any Chinese marque with meaningful volume. The Lexmoto Aura 125 is the most reliable high-volume Lexmoto model.
Before you buy any used bike: Check its specific MOT history at CarHunch.
For example, here's the history for KY66YSM - a 2016 Lexmoto Venom that's failed 4 of its 9 MOTs, with a lighting failure on every single one. Headlamp not working on dipped beam in 2020. Headlamp not working again plus stop lamp failure in 2024. Headlamp out again plus front and rear stop lamps both failed in 2025.
Gets fixed, passes - then fails again with the same class of problem. That pattern in one bike's history lines up with what the aggregate data shows at scale. If you see the same fault class recurring across multiple tests on any used bike - any brand - treat that as a red flag.
A Note on Methodology
On scope: MOT tests check safety-critical systems - lights, brakes, tyres, steering. They do not measure engine durability or every kind of roadside breakdown.
A higher MOT fail rate is not the same thing as "less mechanically reliable overall." It means more items failed the test standard on the days those tests were recorded.
The mix we see here is weighted towards electrical and lighting, not engines or drivetrain. That mix, together with the flat age curve (fail rate little changed from year 1 through year 12), fits early-life and component-level faults more comfortably than a simple story of long-term neglect or pure wear.
This analysis covers 433,244 MOT test records for 230,567 unique Chinese-segment motorcycles and scooters, compared against 2,766,607 tests for 466,765 unique Japanese sub-500cc bikes. Data: DVLA MOT dataset snapshot, January 2026.
On engine capacity and NULL values: LEXMOTO alone has approximately 47% of registrations with no engine capacity recorded - a well-documented quirk of how some smaller importers submit V5C registration data.
For this analysis, Chinese-brand bikes with NULL engine capacity were included alongside confirmed 49-126cc records. These manufacturers have essentially no UK products outside the 50cc-125cc range. This is the only methodology decision that required a judgement call.
On defect categorisation: The DVLA defect text is free text. "Headlamp not working on dipped beam", "nearside headlight inoperative", and "front lamp failure" describe the same fault.
We used MiniLM semantic embeddings (all-MiniLM-L6-v2 via sentence-transformers) to cluster defect descriptions by meaning, then cross-validated those clusters against keyword counts to ensure consistency.
The individual counts in the lighting defect table above reflect all phrasing variants matching each keyword pattern - not a single phrasing.
On advisory validation: We verified directly against the defect_type column in the DVLA data. Advisory share on every named lighting defect type is under 2%. The FAIL type (pre-2018 MOT categorisation) and MAJOR type (post-2018) are equivalent in terms of causing an MOT failure.
On SYM: SYM is Taiwanese, not mainland Chinese. Included because they compete directly in this segment at equivalent price points. Data for SYM and mainland Chinese brands is presented separately throughout.
On CFMOTO and Voge: Both showed implausibly low fail rates (2.1% and 0.1% respectively) due to very small test volumes relative to registration numbers - too new in the UK market. Excluded from reliability comparisons.
Full data tables and the DuckDB query set are available on request. All MOT data is Crown Copyright, used under the Open Government Licence v3.0.
The Bottom Line
The "Chinese bikes are unreliable" narrative is partially true, partially misleading, and entirely missing the most important part of the story.
Yes - a brand-new Chinese 125 is significantly more likely to fail its first MOT than a brand-new Honda or Yamaha. The gap at year one to three is real: 25.8% vs 16.0%.
But the Chinese bike's fail rate stays roughly flat for twelve years. It doesn't fall apart. The Japanese bike climbs to meet it.
The problems are specific - primarily electrics and lighting, not wholesale mechanical disintegration. They show up early in the test history and do not worsen dramatically with age in the aggregate curve.
That profile is what you would look for when talking about early-life faults and infant-mortality-style defects on components (lamps, connectors, brackets) rather than a slow wear cliff. Many of those items are fixable if you catch them before the test.
And within the budget segment, there is genuine variation. The SYM Joyride beats most Japanese bikes on reliability. The JINLUN brand fails 27% of the time. Treating "Chinese bike" as a monolithic category is as useful as treating "Japanese bike" as one.
The data is there - and it is actionable. Before you hand over money for any used 125, enter the registration at CarHunch.
If you see the same fault class recurring across multiple tests, walk away. If the electrical history is clean, you are looking at a lower-risk buy on the dimensions this dataset measures. Every MOT, every advisory, thirty seconds, free.