A used electric car under five years old should generally measure between 88% and 94% State of Health. Between 80% and 88% is normal for an older or higher-kilometre car. Below 80% before 100,000 km, something specific has happened to that vehicle. Below 70%, most manufacturers would replace the battery under warranty — and if yours will not, walk away.
That is the short answer. Everything below is why it is not quite that simple, and why the Australian answer differs from the one you will find everywhere else.
Why you are reading American advice
Search this question and you will get greencars.com, carsa.co.uk, recurrentauto.com — American and British sites, written against American and British warranty terms, in American and British climates.
That matters more than it sounds. Heat is the primary driver of capacity loss in a lithium battery. A pack that lives in Manchester and a pack that lives in Western Sydney do not age the same way, and advice calibrated for one is not calibrated for the other. Australian warranty terms also differ from both — five of the sixteen brands we track publish no capacity guarantee at all in Australia, which changes what a given number means for you.
What State of Health actually is
State of Health is the battery's present usable capacity expressed as a percentage of its capacity when new. A pack at 90% holds nine-tenths of what it originally could.
It is not the charge percentage on your dashboard, which tells you how full the battery is right now. It is not the estimated range, which is a prediction from recent driving efficiency and moves with weather, speed, load and tyre pressure. Those two numbers are visible in every car. State of Health is visible in none of them.
That is the whole reason this question is hard to answer for a car you are standing next to.
What's normal, by age and distance
These are reasonable expectations for a well-treated, actively-cooled electric car in Australian conditions. Individual cars vary, and that variation is exactly the point.
| Age / distance | Typical healthy range | What sits below it |
|---|---|---|
| Under 2 years, under 40,000 km | 93–97% | Early loss is front-loaded; a new EV drops a few per cent quickly, then flattens |
| 2–4 years, 40,000–80,000 km | 90–94% | Below 88% here is worth understanding |
| 4–6 years, 80,000–140,000 km | 86–92% | Below 84% suggests heat, heavy DC charging, or a specific fault |
| 6–8 years, 140,000–200,000 km | 82–89% | Below 80% at this point is genuinely poor |
| 8+ years | 78–88% | Check the warranty has not expired before you need it |
We are being deliberately honest about the limits of this table. Australia has no published, large-scale, independent dataset of measured State of Health by model and age — nobody has built one. These bands are drawn from manufacturer degradation guarantees, observed patterns and the physics of how these packs age, and they are a reasonable frame rather than a measured population. Anyone presenting more precision than this for the Australian market is presenting something they cannot support.
The four bands, and what each should do to a purchase
90–100% — Excellent. Close to original capacity. Buy with confidence on the battery; spend your attention elsewhere on the car.
80–89% — Healthy. Normal wear for age and kilometres. Not a negotiating point on its own. Check the trajectory rather than the number.
70–79% — Watch closely. Meaningful range loss. The car may still suit you perfectly, but it should be priced for its actual capacity, not for the range on the original brochure. Work out what the real range is now and whether it covers your longest regular trip.
Below 70% — Walk-away territory. Under most manufacturers' warranty floor. If the car is still in warranty, this is a claim, not a purchase — and the seller should be making it, not you. If it is out of warranty, you are buying a car that needs a battery.
Why two identical cars differ
Same model, same year, same odometer, several percentage points apart. Three things explain most of it:
Heat. The single biggest factor. A car that has spent five summers parked outside in Penrith has had a harder life than one garaged in Katoomba. Passively-cooled packs — the Nissan LEAF is the notable one in Australia — suffer most, because they have no way to shed heat other than time.
Charging habits. Frequent DC fast charging generates heat and stresses cells, more so on nickel chemistries than on lithium iron phosphate. Routinely sitting at 100% or near 0% is hard on nickel-chemistry packs — though on an LFP car, regular full charges are actually recommended, because the battery management system needs them to recalibrate. The same habit is good practice on one car and bad on another, which is why "how did you charge it?" is only a useful question once you know the chemistry.
Use pattern. Cycles age a battery more than kilometres do. A car doing lots of short trips with frequent top-ups has cycled more than its odometer suggests.
The number that matters more than State of Health
Cell voltage spread. A pack is dozens of cells in series, and the weakest determines behaviour. A pack at 88% with cells tightly grouped is healthy. A pack at 88% with a wide spread has a developing problem that the overall figure conceals entirely.
This is the measurement that separates a real assessment from a number read off a screen, and it is invisible to an owner.
What to do with all this
Establish the car's chemistry and variant, because none of the above means anything until you know what you are looking at. Get the State of Health measured rather than estimated. Look at cell spread, not just the headline figure. Work out the real-world range at the measured capacity — not the claimed range, which for some Australian EVs is overstated by more than 20%. Then check where the warranty stands.
Check any model's battery warranty terms · Work out real-world range at a given State of Health · Book a test — $249
Written for Australian conditions and Australian warranty terms, verified 13 September 2026. The expectation bands above are a reasoned frame, not a measured population dataset — no such dataset is published for Australia, and we would rather say so than imply precision we do not have.