Which Used Cars Have the Most Engine-Failure Complaints? We Counted the NHTSA Data

The news hook is a GM engine investigation. On 2026-08-20 NHTSA opened Engineering Analysis EA26005, covering 997,743 2021-2026 GM trucks and SUVs with the 6.2L L87 V8, after roughly 500 owner reports of engines failing even after the recall repair — some on brand-new replacement engines. That, plus a separate Nissan VC-Turbo thread, has made "which used models actually draw the most engine complaints?" a live buyer question. Rather than echo the recall, we downloaded NHTSA's raw complaint file and computed the answer.

Key facts

• Load-bearing finding: the Ford EcoSport has the highest engine-complaint share of any used model at 86.4% (1,586 of 1,835 complaints); the Kia Soul is second at 69.3%.
• Dataset: NHTSA ODI Complaints flat file (FLAT_CMPL.txt), file dated 2026-08-24, N = 2,237,560 rows = 1,613,687 distinct complaints; 369,158 distinct complaints across model years 2015-2025.
• Metric: per model, the share of that model's distinct complaints (deduped on ODINO) that cite the engine; minimum 300 total complaints per model so tiny-volume models cannot dominate. 188 models clear the threshold.
• Why share, not raw count: the Ford F-150 has the second-most engine complaints (2,208) but only a 19.0% share — that is best-seller volume, not an engine reputation.
• The GM L87 trucks: across all of 2015-2025 they do not top the list, but within the recalled MY2021-2024 window the eight nameplates combined show a 55.0% engine share, up to 82.7% on the Cadillac Escalade.
• News peg: NHTSA EA26005 (opened 2026-08-20), population 997,743; underlying recall 25V-274 (597,630 vehicles, connecting-rod/crankshaft defect).
• Caveat: share is complaint composition, not a failure rate — there is no fleet-size denominator, and NHTSA complaints are unverified.

Which used cars have the most engine-failure complaints?

Ranked by engine-complaint share, the leaders are the Ford EcoSport (86.4%) and a cluster of Kia and Hyundai models built on the Theta II GDI engine — the Kia Soul (69.3%), Kia Optima (61.1%), Hyundai Santa Fe Sport (62.1%) and Hyundai Sonata (52.2%) — plus two Cadillac Escalade variants. None of these are the trucks in the current headlines. We rank by share deliberately: raw counts just rediscover the best-seller list, because a model that sells a million units logs more of every kind of complaint. This mirrors the method behind our companion pieces on which used cars have the most backup camera failure complaints and used cars with the most dashboard display problems.

The full 188-model ranking is in the frozen CSV linked under Sources. The Ford EcoSport figure is a genuine, API-verified outlier — the 2019 model year alone shows 455 complaints citing engine failure and loss of power, tied to the 1.0L EcoBoost — but it rests on a modest 1,835 total complaints and is sensitive to the threshold. If you are cross-shopping Kia and Hyundai models, the same Theta II GDI family also dominates our look at which cars have the most fire complaints.

Do the GM 6.2L L87 trucks or the Nissan Rogue actually top the list?

No — and testing that was the point. The GM 6.2L L87 trucks and the Nissan VC-Turbo Rogue do not lead the market-wide ranking, because high-volume trucks draw large numbers of non-engine complaints that pull their share down. The blended Chevrolet Silverado 1500 sits at 29.6% and the Nissan Rogue at 21.5% across 2015-2025. That is the best-seller confound working exactly as expected; the Ford F-150 has the raw-count silver medal on engine complaints yet only a 19.0% share.

The picture flips once you hold the model year fixed, which is the honest robustness check: are the aggregate leaders just the oldest cars? They are not. In the MY2016 cohort the Kia Soul (71.8%), Hyundai Sonata (63.2%) and Kia Optima (58.5%) lead — the Theta II story. In MY2021 the GM L87 trucks surface strongly: Kia Seltos (71.6%), GMC Yukon (62.3%), GMC Sierra 1500 (59.7%), Chevrolet Silverado 1500 (50.8%). And in MY2023 the Nissan Rogue tops its cohort at 54.9%, with the GM trucks close behind. So the L87 and Nissan signals are real and rise to the top of their own recent cohorts; blending years just dilutes them. Within the recalled window, the trucks are unmistakable:

The L87 defect is a connecting-rod and/or crankshaft manufacturing issue that can cause engine damage and loss of propulsion; recall 25V-274 (owner letters from 2025-06-09) remedies it by inspecting for a diagnostic trouble code, then either switching to higher-viscosity 0W-40 oil or, if the inspection fails, replacing the engine. EA26005 exists precisely because roughly 500 owners report failures after that remedy. The Nissan thread is a separate bearing defect in the VC-Turbo engine, resolved through recalls 25V-437 and 26V-080; our used nissan rogue problems to check before buying piece covers the affected years.

By the numbers

• 1,613,687 distinct complaints in the full file; 369,158 across MY2015-2025.
• 188 models cleared the 300-complaint threshold; 86.4% is the top engine share (Ford EcoSport).
• 55.0% engine share for the eight recalled GM L87 nameplates in MY2021-2024; 82.7% for the Cadillac Escalade.
• 997,743 vehicles in EA26005; 597,630 in recall 25V-274 — different populations, do not conflate.

Based on 2,237,560 NHTSA complaint-component records (1,613,687 distinct complaints) from FLAT_CMPL.txt, file dated 2026-08-24, retrieved and processed 2026-08-25. "Engine" = COMPDESC exactly "ENGINE" or "ENGINE AND ENGINE COOLING" or an "ENGINE AND ENGINE COOLING:ENGINE…" sub-node; cooling, exhaust and emission sub-nodes excluded; powertrain counted separately. Share is complaint composition, not a failure rate.

Does a high engine-complaint share mean a used car will fail?

No. A high engine-complaint share tells you the engine dominates that model's complaint mix, not that the model fails more often than another. There is no fleet-size or mileage denominator here, complaint volume is influenced by media attention and active recalls, and NHTSA does not verify the reports. The share metric strips out the raw best-seller effect but cannot turn an unverified complaint count into a failure rate. A high complaint share is also not the same as a recall — we test how often one leads to the other in do owner complaints predict a car recall. Read the ranking as a shopping-list caution that tells you where to look harder on a given model — then judge the individual car on its own history and a mechanic's inspection.

What a VIN check can and can't tell you here

A VIN check is useful for the presence of a recall and for a car's history, but it cannot prove an engine repair was done. For a specific used GM truck, Rogue, EcoSport or Soul, a Zilocar VIN check flags whether that VIN is named in an open safety recall such as 25V-274 (the same visibility as NHTSA's free tool — presence, not remedy status), and surfaces accident and airbag-deployment records, odometer/rollback checks, theft records, junk and salvage-auction records, ownership history, and past and current sales-listing history with prices, mileage and days-on-market.

Checking a candidate car: run the VIN to see whether it is under the L87 recall at all and to catch accident, salvage, odometer and listing red flags; then confirm the specific engine repair and any investigation exposure through NHTSA's VIN-recall tool and a GM dealer — that remedy status is exactly what a history report cannot prove. A Zilocar VIN check is a fast way to screen recall presence and pull the accident, salvage-auction, odometer, theft, ownership and sales-listing history before you commit; it does not confirm a repair, track EA26005, or classify the title brand.

Method

Source file. NHTSA Office of Defects Investigation consumer complaints flat file, FLAT_CMPL.txt (1,592,308,846 bytes), extracted from https://static.nhtsa.gov/odi/ffdd/cmpl/FLAT_CMPL.zip. The file is last-modified 2026-08-24; we downloaded and processed it on 2026-08-25. It contains complaints filed since 1995-01-01.

Universe and dedup. The file is one row per complaint component, so a single complaint that cites three components appears three times. Raw N = 2,237,560 component rows. Deduplicating on the ODI number (ODINO, field 2) gives 1,613,687 distinct complaints. Every count and share on this page is over distinct complaints, never rows. Restricting to vehicle model years 2015-2025 — the range that defines today's used market — leaves 369,158 distinct complaints.

Grouping. Complaints were grouped by MAKE (field 4), MODEL (field 5) and YEAR (field 6) as the file spells them, with no trim consolidation: the file's ESCALADE and ESCALADE ESV stay separate nameplates, as do SILVERADO 1500 and other cab/box variants where NHTSA records them apart.

What counts as an engine complaint. A complaint counts as engine if its component description (COMPDESC, field 12) is exactly ENGINE (the newer taxonomy), exactly ENGINE AND ENGINE COOLING, or any ENGINE AND ENGINE COOLING:ENGINE… sub-node (the older taxonomy). Cooling, exhaust and emission sub-nodes under the same parent are excluded — a radiator or EGR complaint is not an engine-failure complaint. POWER TRAIN… is counted separately and reported as a second column, never folded into the engine figure. A complaint citing both engine and powertrain is counted once in each column, so the two columns are not additive.

Aggregation and threshold. For each model we report engine complaints ÷ that model's total distinct complaints. Models with fewer than 300 total complaints are dropped, so a nameplate with nine complaints and eight engine complaints cannot take the top spot; 188 models clear the threshold. The reimplementable one-liner is unzip -p FLAT_CMPL.zip | awk -F'\t' '{eng = ($12=="ENGINE" || $12 ~ /^ENGINE AND ENGINE COOLING(:ENGINE|$)/)} …' grouping distinct $2 by $4|$5|$6.

Why share and not raw count. Raw engine-complaint counts simply rediscover the best-seller list — the Ford F-150 is second on raw engine complaints (2,208) at a 19.0% share. Share removes the volume effect but introduces its own: a model that draws few complaints of any other kind will show a high engine share.

Robustness check we ran because it could have reversed the finding. Complaint mix drifts with fleet age, so an aggregate ranking can simply be a ranking of the oldest cars. We therefore recomputed the leaderboard inside fixed single model-year cohorts. It did not reverse: MY2016 is led by the Theta II cars (Kia Soul 71.8%, Hyundai Sonata 63.2%, Kia Optima 58.5%), MY2021 by the GM L87 trucks (Kia Seltos 71.6%, GMC Yukon 62.3%, GMC Sierra 1500 59.7%) and MY2023 by the Nissan Rogue (54.9%). The check did change one claim: it showed the GM and Nissan signals are real inside their own cohorts and are diluted, not absent, in the blended figures — which is why both are reported here rather than dismissed.

Known limits. This is a complaint file, not an incident census: it counts owner reports, which NHTSA does not verify, and there is no sales or fleet-size denominator, so no failure rate per vehicle can be computed and no model here can be called more likely to fail than another. Recall and news publicity measurably raises reporting for the affected vehicles, and the GM and Nissan nameplates in this article are under active recalls and investigations right now, so their recent-cohort figures are the most publicity-exposed numbers on the page. NHTSA's component taxonomy changed over the period, so the old and new engine nodes are unified by the rule above rather than by an official crosswalk. The Ford EcoSport's leading 86.4% rests on a modest 1,835 total complaints and is the figure most sensitive to the 300-complaint threshold. Model years 2024-2025 are still accumulating complaints.

Frozen dataset. The full 188-model ranking — make, model, total complaints, engine complaints, powertrain complaints, engine share and engine-plus-powertrain share — is published as a CSV alongside this article and linked in Sources. The model-year cohort and GM-recall-window tables above are computed from the same file by the same rules.

Related checks

• Used car brands with the most recalls
• What does recall closed mean on a VIN check
• Used Ford Focus problems to check before buying

Sources

• NHTSA ODI Complaints flat file (FLAT_CMPL.zip) — the primary dataset we compute over; file dated 2026-08-24, retrieved 2026-08-25, 2,237,560 rows.
• Zilocar frozen ranking CSV — engine-complaint share by model, MY2015-2025 (188 models, min 300 complaints) — our computed output.
• NHTSA EA26005 Opening Resume (INOA) — Engineering Analysis opened 2026-08-20, population 997,743.
• NHTSA Part 573 Safety Recall Report 25V-274 — GM L87 recall, 597,630 vehicles, connecting-rod/crankshaft defect.
• NHTSA datasets and APIs — file documentation and complaint API used to spot-verify individual model complaint text.
• NHTSA Nissan VC-Turbo recall 26V-080 — secondary thread context (treat aggregator figures as context).
• GM Authority — GM 6.2L L87 post-recall failures, NHTSA EA — news peg.

Last verified: 2026-08-25