How we calculate

Every number on this site comes from a published source. Here are the formulas, the sources, and what each estimate leaves out.

The charging cost calculator

monthly cost = miles per month × (EPA kWh per 100 miles ÷ 100) × your electricity rate

For example: 1,000 miles a month in a car rated 32 kWh per 100 miles, at 18 cents per kWh, is 1,000 × 0.32 × $0.18 = $57.60.

The gas comparison is the same idea: miles ÷ mpg × price per gallon.

Our sources

• EV efficiency: the EPA's official combined city/highway rating, from the vehicle data file on fueleconomy.gov. We use it exactly as published.
• Electricity prices: the EIA's average residential price per state, from Electric Power Monthly. This is an average of everything residential customers actually paid, including fixed charges and taxes.
• Gas prices: the EIA's weekly Gasoline and Diesel Fuel Update. EIA publishes weekly state figures for only nine states; everywhere else we use the national average and say so on the page.

Why charging losses are already included

Charging isn't perfectly efficient. Some energy is lost as heat between the wall and the battery. You don't need to add anything for this, because the EPA measures its rating at the wall outlet. The loss is already in the number.

What the estimate does not include

• Your actual rate plan. State averages hide a lot. Time-of-use and EV-specific plans can charge far less overnight, which is when most people charge. If you know your real rate, type it in; the calculator uses whatever you enter.
• Public fast charging, which usually costs two to three times home charging.
• Cold weather and high speeds. EPA ratings come from standardized tests. Winter driving and sustained highway speeds both use noticeably more energy than the rating suggests.
• Fees and taxes beyond what's already in the EIA average, and any EV registration surcharge your state charges.

The EV finder

The finder scores each car from 0 to 1 on every criterion, multiplies by the percentage you set, and adds the results. Your filters (budget, range, body style, drive, seats, towing, plug) rule cars out before scoring rather than counting against them.

Scores are relative to the cars that passed your filters, not absolute. The longest-range car in that group scores 1 on range and the shortest scores 0, so changing a filter changes the scores. That is deliberate: it ranks the choices actually in front of you.

Where each field comes from

• EPA data, automatic: range, efficiency, body class, drive type and the hours to fill on a 240-volt charger, for every EV of the current model year and newer.
• Entered by hand: starting price, DC fast-charging speed, seats, cargo, tow rating, charging plug, heat pump, and whether the dashboard uses physical controls. EPA publishes none of these, and where a manufacturer publishes a 10-to-80% charging time but no peak kW, we record the time and rank on that instead. One search for a single car's starting price returned four different figures, some including destination charges and some not, which is why these come from the manufacturer's own page with the URL and the date recorded beside the number.
• Reliability: a 1-to-5 score recorded here with the reasoning and a link behind it. Consumer Reports and J.D. Power own their ratings and we do not republish them. Free federal records of complaints, investigations and recalls are at NHTSA.
• Blanks are labelled, not guessed. A car we haven't hand-checked still appears, ranked on its EPA data, but scores a neutral 0.5 on anything unknown and says so on the card. A filter never excludes a car for a value we don't have, so an unchecked car is never silently dropped from your budget.

Why reliability scores deserve extra caution

Reliability data on EVs is thin. Most of these platforms are only a few years old, so nobody has ten-year records on them. Complaint and recall counts are not a fair comparison either, because a car that sold 200,000 units will collect more complaints than one that sold 5,000 regardless of how well it was built, and sales figures that would correct for this are not published in a usable form. Treat the score as an informed opinion with its sources attached, not a measurement.

Charging at home, which outranks everything else

The strongest predictor of whether someone enjoys owning an EV is whether they can plug in where they park overnight. A 240-volt circuit refills almost any EV by morning, which makes fast-charging speed a road-trip concern rather than a daily one. An ordinary household outlet adds roughly 3 to 5 miles of range per hour, so about 40 miles overnight: workable for a short commute, painful otherwise. With no home charging at all, public charging becomes a weekly errand, and fast-charging speed and local charger coverage matter far more than range. Selecting that option raises the weight on charging speed automatically. Get an electrician's quote for a 240-volt circuit before you commit to the car, not after.

Winter

Cold weather cuts range, and the EPA figure does not reflect it. The finder plans with Recurrent's real-world data: across 34 models, EVs averaged 78% of their normal range at 32F, and an earlier study of over 18,000 cars found about 83% with a heat pump and 75% without. So we plan on 17% less range for cars we know have a heat pump, 25% for cars we know don't, and 22% when we haven't checked. Recurrent measures loss against each car's normal real-world range, so applying it to the EPA figure is an approximation. AAA's 2019 lab test found a 41% drop at 20F with the heater running, a harsher worst case for the coldest days. Preconditioning the car while it is still plugged in recovers much of the loss.

Price and range

Prices are the manufacturer's starting MSRP before destination fees, which usually add over $1,000. The starting price buys the base trim, and we don't yet record which EPA range belongs to which trim, so the finder ranks every car on the lowest EPA range among its trims and wheels and shows the full span. By default the finder only shows cars with a checked price, so your budget always holds; you can choose to include unpriced cars, which then score neutral on price.

Incentives

The federal purchase credits for new (up to $7,500) and used (up to $4,000) clean vehicles ended for vehicles acquired after September 30, 2025, and the federal home charger credit ended for equipment placed in service after June 30, 2026. There is no federal EV purchase credit for a 2026 buyer. Some states and utilities still offer rebates, and those change without much notice, so check your state energy office and your own utility. Tax rules change; confirm with the IRS or a tax professional before it affects a purchase.

What the finder still does not know

Insurance costs, how far the nearest dealer's service department is, whether the car is even sold in your state, lease deals (often better value than the sticker suggests), and how a car feels to drive. Use the shortlist to decide what to test drive, not what to buy.

The carbon payback calculator

payback miles = battery footprint (g) ÷ (gas CO2 per mile − EV CO2e per mile)

The battery footprint is battery size (kWh) × a production factor (kg CO2e per kWh). The EV's per-mile figure is its EPA efficiency (kWh per 100 miles) × the emissions of the electricity it charges from. The gas car's figure is 8,887 g CO2 per gallon ÷ its mpg.

Manufacturing: a range, not one number

There is no official per-model figure for the carbon cost of building an EV. We use the ICCT's July 2025 table of battery production emissions, itself built on Argonne National Laboratory's GREET 2024 model. It varies by chemistry and by where the cells were made: 52 to 80 kg CO2e per kWh across LFP, NMC 622 and NMC 811 cells from Europe, the United States, South Korea, Japan and China. You choose what you know. If you know nothing about your battery, we use 72.8 kg CO2e per kWh, the ICCT's sales-weighted figure for NMC 622 batteries in cars sold in the EU in 2024, and say so on the page. Other published estimates run from 37 to 115 kg CO2e per kWh, and the harmonized review by Peiseler et al. gives a median of 74 for NMC 811. Move the number if you have a better one; the payback result scales with it.

We treat the battery as the whole of the EV's extra manufacturing footprint and assume the rest of the vehicle costs roughly what a gas car does to build. The ICCT estimates BEV production emissions are about 40% higher than a gas car's, with the battery as the reason.

Battery size isn't in EPA's data. We estimate it as EPA range × EPA efficiency, the energy drawn from the wall to cover the rated range. Because that includes charging losses, the estimate usually overstates the pack by roughly 10 to 15% (for example, it gives 99 kWh for a 2026 Equinox EV FWD, whose maker lists 85), so the payback miles shown are conservative. Enter your real battery size if you know it.

Electricity: your state, or a specific source

For grid charging we use EPA eGRID2023's CO2e output rate for electricity generated in your state, then add eGRID's 4.2% grid loss so the figure reflects delivered electricity. These rates cover emissions at the power plant. They describe generation within the state, not imports, so states that buy a lot of power from neighbors are approximated. EPA's own Power Profiler gives a ZIP-code figure if you want one.

If you choose a specific source (solar, wind, nuclear, hydro, geothermal, gas, coal), we use the IPCC's lifecycle median for that source. Lifecycle figures include building the plant and supplying its fuel, which the grid figure does not, so the source options count a little more than a grid figure built the same way would. Rooftop solar gets no grid-loss adjustment; every other source does.

Gasoline: tailpipe only

We use EPA's 8,887 g CO2 per gallon, which matches the tailpipe figures on EPA's own vehicle labels. We do not add emissions from extracting, refining and delivering the fuel. That omission favors the gas car, so the payback miles shown are, if anything, slightly too high.

Trees: an illustration, not an offset

EPA's equivalencies calculator says a medium-growth urban tree seedling absorbs about 60 kg of CO2 over its first 10 years. We divide the annual saving by that figure. The point is scale, not accounting: the EV's saving happens this year, while a seedling takes a decade to absorb the same amount, and only if it survives. Don't read the tree number as something you've offset.

What the payback estimate does not include

• Battery replacement, recycling credits, or end-of-life treatment for either vehicle.
• The grid getting cleaner over time, which shortens a real payback.
• Cold weather and highway speeds, which raise real-world EV consumption above the EPA rating.
• Tire and brake wear, maintenance, or anything other than CO2-equivalent greenhouse gases.

How often we update

Monthly for prices and vehicles. EIA publishes new state electricity prices near the end of each month, and we refresh the EPA vehicle list at the same time. Every figure on the calculators shows the date of the data it came from. eGRID is annual (EPA released 2023 data in 2025) and the battery and IPCC factors change only when the underlying studies do.

Found an error?

Tell us which figure and which source, and we'll check it. Contact us.