The Reliability grade, explained

Reliability is the subject people most want a number for and the one where no honest number exists. Nobody measures dependability the way a dynamometer measures torque; what exists is predictions, surveys, recall records and reputations, and a grade has to be built from those. MotorGPA’s Reliability grade is exactly that: eight judged criteria, every one scored against written guidance and published with its reasoning and its source, weighted toward the evidence that has the best track record of being right.

It is also the joint-heaviest subject in the overall GPA. That is a statement of priorities: a car that is agreeable but unreliable is a car you regret, and the weighting encodes it.

What the criteria measure

Three groups. Track record is the core and carries most of the weight: predicted reliability (the published predictions from the large owner surveys, weighted ×3), the brand’s reliability history, the model’s recall and technical service bulletin record, and the powertrain’s durability reputation. Build quality is what you can see and touch: interior material quality and fit and finish. Service and support is what happens when something does go wrong — the dealer and service network, and parts availability and repair cost. Every one of the eight is judged; there is no measured criterion in this subject, which is why the reasoning beside each score matters more here than anywhere else.

How the fleet grades

Reliability grades cluster tighter than most subjects, because the judged guidance reserves the top of the scale for brands with a long verified record and the bottom for models with a documented pattern of failures, and most cars are neither. The heaviest criterion, predicted reliability, medians at 65 across the fleet — the survey data it rests on is itself centred, and the grade inherits that.

Who leads, and why

The clearest pattern in this subject is by powertrain. Hybrids median 81.4 points (A-) and electric vehicles 58.5 (C-), the widest spread of any subject by class. The reason is what the criteria reward: a long verified record. The hybrid fleet is led by nameplates with two decades of the same powertrain architecture, while most EVs on sale are first- or second-generation products from brands whose survey history is short — and the predicted-reliability surveys, which carry the most weight, penalize newness and software faults heavily. Why hybrids out-grade EVs on Reliability takes the question apart class by class.

How to read the grade

A Reliability grade is a synthesis, and it should be read as one. An A means the published predictions, the recall record, the build quality and the support network all point the same way. A C usually means they disagree — a strong brand record with a weak first-year prediction, say — and the criterion-level scores tell you which. A D or F is rare and means the evidence is consistently bad.

Two cautions. First, the grade is a prediction about a model year, not a guarantee about a car; a specific used example has a history the survey does not know. Second, brand reputation is one of eight criteria, not the grade — a model can out-grade or under-grade its badge, and several do, which is what the leaders and laggards tables above are showing.