Cars With the Best Gas Mileage in 2026: the EPA’s Own Numbers, and Where They Stop Being True

The MPG figure on a new car’s window sticker is not measured by the government. The manufacturer runs the test — usually on a pre-production prototype — and reports the result to the Environmental Protection Agency, which re-runs about 15% to 20% of them at its own laboratory to check the work. That is not a leak or an accusation. It is the second paragraph of the EPA’s page explaining how vehicles are tested.

Which is worth knowing before reading any list of the cars with the best gas mileage, including this one. The rankings below come from the EPA’s own 2026 data, because every other list you will find is a repackaging of the same source. What almost nobody adds is the part that decides whether the number applies to you: who ran the test, at what speed, with how much weight in the car, and what the agency’s own drivers report when they fill up in the real world.

So: the lists first, then the fine print that changes how to read them.

The most fuel-efficient 2026 cars, by EPA class

The EPA sorts vehicles into size classes and publishes the leader in each. Two tables, because mixing electricity and gasoline in one ranking produces a list where nothing is comparable to anything.

Here is the top of each class when electric-drive vehicles are included. The numbers for electric vehicles are MPGe — miles per gallon of gasoline equivalent, a conversion of kilowatt-hours into the energy content of a gallon of fuel, since electricity is not sold in gallons.

EPA class	2026 leader	Combined MPG(e)
Two-seaters	Bugatti Rimac Nevera R (EV)	51
Minicompacts	Fiat 500e (EV)	116
Subcompacts	Honda Prelude (hybrid)	44
Compacts	Mercedes-Benz CLA250 Plus with EQ Tech (EV)	126
Midsize cars	Tesla Model 3 Standard RWD (EV)	139
Large cars	Lucid Air Pure RWD, 19-inch wheels (EV)	146
Small station wagons	Nissan LEAF 75 kWh (EV)	121
Midsize station wagons	Mercedes-Benz EQE 320 Plus SUV (EV)	93

And here is the same list restricted to vehicles that actually burn fuel — the one most people are looking for when they type the question.

EPA class	2026 leader (no electric drive)	Combined MPG	EPA annual fuel cost
Two-seaters	Mazda MX-5 (manual or automatic)	29	$2,600
Minicompacts	MINI Cooper C / S / JCW Convertible	30	—
Subcompacts	Honda Prelude (hybrid)	44	$1,400
Compacts	Toyota Corolla Hybrid	50	$1,250
Midsize cars	Toyota Prius (hybrid)	57	$1,100
Large cars	Hyundai Sonata Hybrid Blue	51	$1,200
Small station wagons	Toyota Crown Signia AWD (hybrid)	38	$1,600
Midsize station wagons	Honda CR-V FWD (hybrid)	40	$1,550

Every single class leader that burns fuel is a hybrid, except the two sports-car classes where nobody is shopping on fuel economy anyway. The 2026 Prius, at 57 mpg combined, is the most efficient thing on the American market that runs purely on gasoline.

Two details in that first table are worth a second look. The most efficient two-seater in the United States is a multi-million-dollar electric hypercar rated at 51 MPGe — a worse figure than a Sonata Hybrid, because the conversion penalises a car built for 1,900 horsepower. And the subcompact class is led by a hybrid coupé at 44 mpg, not by an electric car: the leader board is not a straight line from gasoline to electric.

The SUVs and crossovers with the best gas mileage in 2026

SUVs get their own EPA classes, and they are where the market actually is — more on that below.

EPA class	Including electric drive	MPG(e)	Fuel-burning leader	Combined MPG
Small SUVs	Tesla Model Y Standard RWD, 18-inch (EV)	138	Kia Niro FE (hybrid)	53
Standard SUVs	Audi Q4 45 e-tron (EV)	115	Jeep Cherokee AWD (hybrid)	37
Minivans	Chrysler Pacifica Hybrid (PHEV)	48	Toyota Sienna 2WD (hybrid)	36

A Kia Niro FE at 53 mpg combined is within four of the Prius, in a body that carries a family. That is the single most under-appreciated number on this page: the fuel-economy penalty for buying a small SUV instead of a compact hybrid sedan has almost disappeared, provided the SUV is a hybrid.

The gap inside the same nameplate is larger than the gap between nameplates. A Honda CR-V FWD with the 2.0-litre hybrid system is rated 40 mpg combined and $1,550 a year; the same CR-V FWD with the 1.5-litre turbo is 30 mpg and $2,050. Same badge, same body, $500 a year apart.

At the other end, the least efficient standard SUV of 2026 is a supercharged 6.2-litre Dodge Durango AWD at 13 mpg combined, which the EPA costs at $5,850 a year in premium fuel.

Pickups, and the vehicles with no rating at all

EPA class	Including electric drive	MPG(e)	Fuel-burning leader	Combined MPG
Small pickups	Ford Maverick HEV FWD (hybrid)	38	Ford Maverick HEV FWD	38
Standard pickups	Rivian R1T Dual Max, 22-inch (EV)	87	Chevrolet Silverado 2WD / GMC Sierra 2WD, 3.0L diesel	25

The Maverick hybrid is the only vehicle in these tables that leads its class both with and without electric drive included: at 38 mpg it beats every electric small pickup, because there isn’t one.

Now the part that ruins the tidiness of the table. Heavy-duty pickups, large vans and the biggest SUVs have no EPA fuel economy rating at all. Manufacturers are not required to test vehicles above a gross vehicle weight rating of 8,500 pounds — for pickups, above 8,500 pounds with a bed of 72 inches or more; for cargo vans, above 8,500 pounds; for passenger SUVs and vans, at 10,000 pounds and up. The EPA lists the Chevrolet Silverado 2500/3500, the Ford F-250/350, the RAM 2500/3500 and the GMC Sierra 2500/3500 as examples. If you are cross-shopping a three-quarter-ton truck and cannot find its MPG anywhere, that is why: the number does not exist.

And one more line from the same table that undoes the whole premise of ranking by MPG. The Silverado 2WD with the 3.0-litre diesel is rated 25 mpg combined and $3,200 a year. The Silverado 2WD with the 2.7-litre gasoline turbo is rated 20 mpg combined and $3,100 a year. Five miles per gallon better, and it costs more to run, because the EPA’s guide assumes $5.21 a gallon for diesel against $4.15 for regular. Miles per gallon is not the bill.

Who runs the test, and what it looks like

Fuel economy is measured in a laboratory, on a dynamometer, following cycles specified by federal law. Manufacturers test their own vehicles — typically pre-production prototypes — and send the results to the EPA, which confirms roughly 15% to 20% of them by re-testing at its National Vehicles and Fuel Emissions Laboratory. They are also not required to test every version on sale: one representative vehicle per combination of loaded weight class, transmission class and basic engine.

Since 2008 there have been five cycles, not two. Here is what each one actually does — and the reason this table matters is that most of these figures are lower than what people picture when they read the word “highway”.

	City	Highway	High speed	A/C	Cold temp
Top speed	56 mph	60 mph	80 mph	54.8 mph	56 mph
Average speed	21.2 mph	48.3 mph	48.4 mph	21.2 mph	21.2 mph
Distance	11 mi	10.3 mi	8 mi	3.6 mi	11 mi
Duration	31.2 min	12.75 min	9.9 min	9.9 min	31.2 min
Stops	23	none	4	5	23
Idling	18% of time	none	7%	19%	18%
Engine start	cold	warm	warm	warm	cold

The highway cycle never exceeds 60 mph and averages 48.3 mph. The high-speed cycle tops out at 80. The air-conditioning cycle simulates 95°F with a sun load. The combined figure on the sticker is not an average of your driving: it is 55% of the city number plus 45% of the highway number, a fixed weighting that has nothing to do with your commute.

Then there is what the test assumes about the vehicle itself:

• Four-wheel-drive vehicles are tested in two-wheel drive. Engaging all four wheels makes the engine work harder and adds transfer-case and differential losses that the rating never saw.
• The vehicle is assumed to carry 300 pounds of passengers and cargo. Two adults and a weekend bag put you over it.
• The test assumes flat ground. No hills, no mountain passes.
• Roof boxes, bike racks and trailers are not modelled at all.

None of this makes the ratings dishonest. They are a comparison tool, run identically for every vehicle so that the ranking between two cars means something. They were never a forecast of what your tank will do.

What the EPA’s own drivers report

Here is the part almost no list mentions: the same agency that publishes the sticker numbers also publishes what drivers report. Its My MPG feature lets owners log fill-ups, and the resulting averages sit on fueleconomy.gov next to the official rating.

The current top ten by driver-reported average, restricted to vehicles with records from at least ten drivers:

Rank	Vehicle	Driver average	EPA combined	Drivers
1	Honda Insight, 1.0L 3-cyl manual (2004–2006)	71.6	52	16
2	Toyota Prius (2016–2022)	57.0	52	44
3	Hyundai Ioniq Blue (2017–2020)	56.8	58	10
4	Toyota Corolla Hybrid (2020–2022)	53.6	52	10
5	Honda Insight (2019–2022)	53.4	52	11
6	Hyundai Ioniq (2017–2020)	51.9	55	18
7	Toyota Prius c (2012–2016)	51.2	50	70
8	Geo Metro XFI (1990–1994)	50.8	47	23
9	Kia Niro (2017–2019)	50.3	49	22
10	Honda Civic CRX HF (1990–1991)	49.8	43	13

Four things fall out of that table.

The best real-world average any group of ten or more drivers has posted, 71.6 mpg, belongs to a two-seat aluminium hybrid discontinued twenty years ago, and it beats its own EPA rating by 38%. Nothing in the list is newer than a 2022 model year. Two entries are from 1990.

Two of the ten came in below their sticker — the Ioniq Blue by 1.2 mpg and the Ioniq by 3.1. So the drift is not a uniform “the sticker lies high” story. It runs both ways, and which way it runs for a given car is not predictable from the rating.

The spread between drivers of the same model is wider than the spread between models. Somebody in that Prius c sample is getting 40-something, somebody else 60-something, on the same rating.

And the honest caveat, which the EPA prints itself: these are self-selected owners who bothered to log fill-ups, on samples between ten and seventy people. It is not a controlled study. It is exactly what a controlled study cannot produce — the number a real person got, on real roads, with a real right foot.

The pattern in the cars people actually own

The top-ten list is built from record-holders. The same database also holds ordinary vehicles, and it is queryable, so I pulled the model years with the largest samples for a handful of high-volume nameplates. What comes out is more useful than the podium.

Vehicle	EPA combined	Driver average	Drivers	Reported range
2015 Ford F-150 4WD, 3.5L turbo	19	16.5	11	12–23
2015 Honda Civic, 1.8L	33	29.9	10	21–39
2015 Subaru Outback AWD, 2.5L	28	25.6	28	20–31
2016 Subaru Outback AWD, 2.5L	28	25.5	19	19–30
2017 Subaru Outback AWD, 2.5L	28	26.3	14	17–32
2017 Honda CR-V AWD, 1.5L	29	29.9	13	22–35
2019 Toyota RAV4 Hybrid AWD	40	39.2	15	34–46
2016 Toyota Prius	52	60.1	20	51–86
2017 Toyota Prius	52	55.2	13	45–64

The direction is not random. Every hybrid in that sample lands at or above its rating; the all-wheel-drive wagon lands 6% to 9% below it across three consecutive model years, and the turbocharged pickup 13% below. That fits what the test procedure actually does: four-wheel-drive vehicles are rated in two-wheel drive, and a small turbo engine hits its efficient operating point on a gentle laboratory cycle far more often than in a driveway.

One more thing that table teaches. Plug-in hybrids are excluded from the EPA’s ranking, and the raw data shows why: the 2017 Prius Prime sample averages 100.3 mpg and the 2018 one 145.0, with individual records above 250. Those are not fuel economy figures, they are the arithmetic of dividing miles by the gasoline you did not burn because the car was plugged in.

If you want to run the same measurement on your own car, the arithmetic and the traps are here, and for electric vehicles the equivalent metric is miles per kWh, which is worth understanding before comparing a 138 MPGe crossover to a 53 mpg hybrid.

The MPG illusion: why the biggest saving is at the bottom of the table

Miles per gallon is a ratio with the useful quantity in the denominator, and that makes it deeply misleading for comparing improvements. The EPA states the problem in the notes to the window sticker itself.

Over 1,000 miles, moving from 10 mpg to 15 mpg saves about 33 gallons. Over the same 1,000 miles, moving from 30 mpg to 35 mpg saves about 5 gallons. Same five-MPG step, nearly seven times the fuel.

Scaled to the EPA’s assumed 15,000 miles a year, using its own 2026 figures:

Swap	Gallons per year	Saved
Ram 1500 6.2L supercharged (12 mpg) → Silverado 2WD 2.7L (20 mpg)	1,250 → 750	500 gallons
Toyota Corolla Hybrid (50 mpg) → Toyota Prius (57 mpg)	300 → 263	37 gallons

Trading the thirstiest pickup for an ordinary four-cylinder one saves thirteen times more fuel than trading the second-best hybrid in America for the best one. This is why the EPA also prints a consumption figure — gallons per 100 miles — next to the MPG on the label: consumption is linear, economy is not.

The practical version: if a household has two vehicles, the fuel saving lives in whichever one has the lower MPG, not in replacing the efficient one with something marginally more efficient.

The dollar figure on the sticker assumes you live nowhere in particular

The annual fuel cost printed on every window sticker rests on two assumptions: 15,000 miles a year, split 55% city and 45% highway, and a projected national fuel price. The 2026 Fuel Economy Guide uses $4.15 a gallon for regular, $5.14 for premium, $5.21 for diesel and $0.15 per kWh for electricity.

Against reality, the national assumption holds up reasonably well. In the week ending 10 August 2026 the average retail price of regular gasoline in the United States was $4.006 a gallon and on-highway diesel $5.257 — within a few cents of the guide’s numbers, though regular is up 89 cents from a year earlier.

The regional spread is what breaks it. In that same week, regular averaged $3.543 on the Gulf Coast and $5.075 on the West Coast — a difference of $1.53 a gallon.

Run that through 15,000 miles in a 30 mpg car, which needs 500 gallons a year. On the Gulf Coast that is about $1,772. On the West Coast, about $2,538. The same car, the same driving, $766 apart.

For comparison, improving from 30 mpg to 40 mpg at the national average price saves about $501 a year. Moving from Houston to Los Angeles costs more than a ten-MPG upgrade earns. The sticker cannot know that, and neither can any list of the most fuel-efficient cars, including this one.

What actually moves your number

The EPA publishes ranges for the things a driver controls, each tied to a study rather than to folklore. These are the honest ones:

Change	Effect on fuel economy
Aggressive driving (speeding, hard acceleration and braking)	−15% to −30% at highway speeds; −10% to −40% in stop-and-go
Driving over 60 mph	roughly −7% per additional 5 mph
Large, blunt rooftop cargo box	−2% to −8% city, −6% to −17% highway, −10% to −25% at 65–75 mph
Rear-mounted cargo box or tray	−1% to −2% city, −1% to −5% highway
Every extra 100 pounds carried	about −1%
Air conditioning on maximum	−5% to −25%
Gasoline with 10% ethanol	−3% to −4%
Winter-blend fuel versus summer	about 1.7% less energy per gallon
Tires below recommended pressure	up to −0.6% recovered by inflating correctly
Wrong grade of motor oil	−1% to −2%
Engine not yet broken in (first 3,000–5,000 miles)	below the vehicle’s eventual best

Two entries there deserve emphasis, because they are the only ones that are free.

Idling gets zero miles per gallon and burns a quarter to half a gallon an hour. Restarting a modern engine costs about ten seconds’ worth of fuel, so any stop longer than a minute is cheaper with the engine off — the EPA’s own guidance, based on an Argonne National Laboratory study of starter wear, is to keep it under roughly ten restarts a day and drive at least five miles between them to let the battery recover.

And the one that explains why this whole article exists: the EPA reports that driver feedback devices improve fuel economy by about 3% for the average driver, and about 10% for drivers actively using them to save fuel, citing a large field test run by the Institute of Transportation Studies at UC Davis. Measurement itself is worth more than most of the mechanical tweaks on that list.

Fleet-wide, the record is real — and electricity is carrying it

The EPA’s Automotive Trends Report, updated 19 February 2026, gives the national picture for model year 2024: average new-vehicle real-world fuel economy reached a record 27.2 mpg, up 0.1 from the previous year, improving in 16 of the last 20 years and up 41% since model year 2004.

Three qualifications sit inside that record, and they are all in the same report.

Without battery-electric and plug-in hybrid vehicles, the model year 2024 average would have been 1.7 mpg lower. Over the five years to 2024, thirteen of the fourteen largest manufacturers improved — but seven of them show declining fuel economy once BEVs and PHEVs are excluded. Toyota is the clearest exception, gaining 3.3 mpg with or without them, largely by building more conventional hybrids.

The market keeps moving toward the heavier classes. In model year 2024, 66% of new vehicles were trucks under the regulatory definition, and truck SUVs alone accounted for half of all production. That class did improve, to a record 25.7 mpg. Sedans, wagons and car SUVs went the other way.

The spread between manufacturers is enormous. For model year 2024, Tesla’s all-electric fleet averaged 117.1 mpg, Honda 31.0, Hyundai 29.8 and Kia 29.2, while Stellantis came last among the large manufacturers at 22.8, behind GM at 22.9 and Ford at 23.4.

Which is a useful frame for any “best gas mileage” list: the difference between a good badge and a bad badge, fleet-wide, is about 8 mpg. The difference between a careful driver and an aggressive one, on the same car, can be larger than that.

Where Magica fits, honestly

I build a mileage and vehicle app, so let me be exact about what it does and does not do here. It does not predict fuel economy, it does not rate cars, and it will not tell you which SUV to buy.

What it does is the measurement the EPA cannot run for you. Magica keeps a fill-up log — litres or gallons, cost, odometer reading, station — and turns it into your actual consumption trend and average efficiency over time, for combustion and for electric charging alike. Receipts get scanned and stored against the entry. The consumption analysis is per vehicle, so two cars in one household stay separate, and the whole record exports as CSV or PDF.

Two things about that matter more than the feature list. Everything stays on the device: there is no account, and no server of mine holds a log of where you buy fuel. And the number it produces is the only fuel-economy figure that describes your car rather than a prototype of it — which, going by the EPA’s own field data, is worth about 3% to 10% on its own, simply because you can see it.

So which one should you actually buy?

If the question is strictly which vehicles use the least fuel in 2026, the EPA’s answer is unambiguous and sits in the tables above: a Prius at 57, a Niro FE at 53, a Sonata Hybrid Blue at 51, a Corolla Hybrid at 50, a Maverick hybrid at 38 in a pickup body, a Silverado diesel at 25 in a full-size one.

If the question is where the money is, three corrections apply. The saving lives at the bottom of the MPG scale, not the top. The fuel price where you live can outweigh a ten-MPG improvement. And the version of the model matters more than the badge — the same CR-V is 40 mpg or 30 mpg depending on which engine is under the hood.

If you are buying used, the sticker is a decade-old test of a prototype, and what to actually check on a used car will tell you more than its original rating. If you are weighing brands, the reliability rankings disagree with each other in ways worth understanding first. And once the car is yours, the figure that predicts your next five years is not MPG at all — it is total cost per mile, of which fuel is only one line, alongside the service record you keep.

Frequently asked questions

Which car has the best gas mileage in 2026?

Among vehicles that burn fuel, the Toyota Prius, at 57 mpg combined on the EPA’s 2026 ratings. Counting electric drive, the Lucid Air Pure RWD with 19-inch wheels leads at 146 MPGe, but MPGe is a conversion of kilowatt-hours into the energy content of a gallon of gasoline, not a measure of fuel burned.

Which SUV has the best gas mileage in 2026?

The Kia Niro FE, at 53 mpg combined, is the most fuel-efficient small SUV that runs on fuel; the Jeep Cherokee AWD hybrid leads the standard SUV class at 37. With electric drive included, the Tesla Model Y Standard RWD on 18-inch wheels leads small SUVs at 138 MPGe and the Audi Q4 45 e-tron leads standard SUVs at 115.

Which pickup truck has the best gas mileage?

The Ford Maverick HEV FWD at 38 mpg combined among small pickups, and the Chevrolet Silverado 2WD or GMC Sierra 2WD with the 3.0-litre diesel at 25 mpg among standard pickups. The electric Rivian R1T Dual Max is rated 87 MPGe. Heavy-duty pickups above 8,500 pounds gross vehicle weight rating have no EPA rating at all.

Are EPA MPG ratings accurate?

They are accurate as a comparison between vehicles, because every vehicle runs the same five laboratory cycles. They are not a forecast of an individual result. The manufacturer runs the test on a representative pre-production vehicle and the EPA re-tests about 15% to 20%; the car is tested in two-wheel drive if it has four, with 300 pounds of load, on simulated flat ground, and the highway cycle never exceeds 60 mph.

Why is my real MPG lower than the sticker?

Usually a combination of speed, temperature and trip length. Driving above 60 mph costs roughly 7% per additional 5 mph, aggressive driving costs 10% to 40% in traffic, a rooftop box costs up to 25% at interstate speeds, air conditioning on maximum up to 25%, and 10% ethanol fuel 3% to 4%. Short trips from a cold start are the worst case, because the engine spends most of the trip below its efficient temperature.

What is MPGe and can I compare it to MPG?

MPGe is miles per gallon of gasoline equivalent: the distance a vehicle travels on an amount of electricity holding the same energy as a gallon of gasoline. It allows a rough comparison across fuels but not a cost comparison, since the guide’s assumed price is $4.15 a gallon of regular against $0.15 per kWh of electricity.

How much does gas mileage actually save per year?

Less than the MPG gap suggests at the efficient end, more at the thirsty end. Over 15,000 miles, going from 12 mpg to 20 mpg saves 500 gallons a year; going from 50 mpg to 57 mpg saves 37. Regional fuel prices matter as much: in the week ending 10 August 2026 regular gasoline averaged $3.543 a gallon on the Gulf Coast and $5.075 on the West Coast.

Is a diesel cheaper to run than a gasoline engine?

Not automatically. The EPA rates the 2026 Silverado 2WD 3.0-litre diesel at 25 mpg combined and $3,200 a year, and the 2.7-litre gasoline version at 20 mpg and $3,100 — five miles per gallon better and slightly more expensive to fuel, because the guide assumes diesel costs $5.21 a gallon against $4.15 for regular.

Sources

• U.S. EPA / U.S. Department of Energy, fueleconomy.gov: 2026 Most and Least Efficient Cars; 2026 Most and Least Fuel Efficient Trucks, Vans and SUVs; Top Ten (driver estimates); How Vehicles Are Tested; Which Vehicles Are Tested; Detailed Test Information; Many Factors Affect Fuel Economy; Driving More Efficiently; Gasoline Vehicles: Learn More About the Label. All accessed 18 August 2026.
• U.S. EPA / U.S. DOE, Fuel Economy Guide 2026 (PDF), pages on fuel economy leaders, annual fuel cost assumptions and fuel-saving technologies.
• U.S. EPA, Highlights of the Automotive Trends Report, last updated 19 February 2026 (model year 2024 data).
• U.S. Energy Information Administration, Gasoline and Diesel Fuel Update, release of 11 August 2026 (week ending 10 August 2026).
• Studies cited by fueleconomy.gov for driver-behaviour effects: Oak Ridge National Laboratory (speed–MPG and cargo/tire-pressure testing), Argonne National Laboratory (stop-start effects on starting systems), Institute of Transportation Studies, UC Davis (ECODRIVE I-80 fuel economy feedback field test).

Fuel economy figures are EPA estimates for model year 2026 as published at the date above and can change when the agency updates its data; fuel prices are weekly national and regional averages and will not match your local pump.