Some links on our site may earn us a commission at no extra cost to you. Learn more.
Updated August 2026: each carrier’s “airline approved” marketing checked against the airline’s own published under-seat maximum — it doesn’t hold up for two of the four. Petmate’s own page rates the Two Door Top Load to 15 pounds. All four carriers re-priced direct from each maker.
Buy the Sherpa Original Deluxe, $69.99, if one carrier has to do both jobs: it’s rated to 22 pounds, its spring-wire frame compresses to fit under a seat, and Sherpa backs it with a refund if an airline denies boarding. The Sleepypod Air, $214.99, is the one carrier here independently crash-tested for pets up to 18 pounds — buy it if you fly more than once a year. The Petmate Two Door Top Load, $64.99, is the vet-trip specialist and, by its own published dimensions, cannot fly with you. And the SturdiBag XL, roughly $109, carries 40 pounds of cat but its own maker rules it out of the cabin entirely.
Prices are approximate, current as of publication, and change frequently — confirm the latest price with the retailer before buying.
Check the Airline’s Own Numbers Before You Buy
Every carrier here markets itself around flying, one way or another. Only one airline’s numbers matter once you’ve bought the bag: the one you’re actually booked on. United publishes an 18 x 11 x 11 inch maximum for a soft-sided in-cabin carrier and charges $150 each way for the pet, per its own pet policy, current as of July 2026. Line that up against what each maker publishes about its own product and two of these four carriers fail before you ever reach the gate.
Carrier
The marketing claim
The maker’s own numbers
Verdict against United’s 18x11x11 max
Sherpa Original Deluxe
“Airline Approved,” Guaranteed On Board program
19 x 11.8 x 11.5 in at rest (Large); spring-wire frame compresses further
Holds up — a hair over at rest, built to compress into the window
Sleepypod Air
“Largest carrier approved for in-cabin air travel”
16 in compressed length x 10.5 x 10.5 in, per Sleepypod’s own spec page
Holds up with room to spare on every axis
Petmate Two Door Top Load
Sold as a “travel kennel”
24 x 16.8 x 14.5 in exterior, rigid hard plastic, per Petmate’s own page
Does not hold up — every dimension exceeds the max, and plastic doesn’t compress
SturdiBag XL
“Flex-height design…to fit under airline seats” (SturdiBag line marketing)
Sturdi Products’ own size chart puts XL at 40 lb capacity; its own overview page states XL and XXL “are too large for in-cabin use”
Does not hold up at this size — ruled out by the maker’s own page
Read that third row twice. Petmate never claims cabin approval outright, but it’s shelved and searched alongside “travel” carriers, and nothing on its product page tells you it can’t fly. The dimensions do that instead, if you do the comparison yourself. Same story with the SturdiBag line’s flex-height pitch: real for the Small, Medium, and Large sizes, but Sturdi’s own overview page draws a line at XL. Buy the big one for a big cat’s car rides and vet trips. Don’t buy it expecting it to fly.
Whichever carrier you buy, the size that matters is the one printed on your airline’s own policy page — not the one on the carrier’s box.
Buy the Sherpa for Everyday Flying
The Original Deluxe is what most cats should ride in, and the numbers back the reputation. Its spring-wire frame lets the rear of the bag flex down several inches to slide under a low seat, then spring back out at the destination. Sherpa’s Guaranteed On Board program refunds your pet fee and ticket if the carrier is denied boarding when used within the program’s terms — the only carrier here backed by a named promise like that rather than a standard return policy.
The recurring complaint across retailer reviews is zipper durability under a determined cat. That’s a real issue for an escape artist, not a design flaw for most cats.
Skip this if your cat tops 22 pounds, or treats zippers as a puzzle to solve.
Pay the Sleepypod Premium Only If You Fly Often
The Air earns its price with the one credential nothing else here has: independent crash testing by the Center for Pet Safety, rated for pets up to 18 pounds. It compresses from a 22-inch expanded length down to 16 inches, comfortably inside United’s 18-inch cap, and doubles as a car seat with a seatbelt routing.
At $214.99, it’s the most expensive carrier in this comparison by a wide margin, and its 18-pound cap is the lowest of the four — a large cat simply won’t fit inside a crash-tested safety margin built for a smaller animal.
Skip this if your cat flies less than once a year, or weighs more than 18 pounds. The Sherpa does the same job for a third of the price at that frequency.
Buy the Petmate Only for the Clinic
Ask a vet tech what the actual fight is on exam day and it’s rarely the cat getting to the clinic — it’s getting the cat out of the carrier. The Petmate’s top-loading steel door solves that directly: lower the cat in vertically, and the whole top half unlatches for an exam without the cat leaving the base. Hard plastic wipes clean after accidents, and at $64.99 it’s the second-cheapest carrier here.
Petmate’s own page also rates it to 15 pounds — worth knowing if you’re buying for a bigger cat.
Skip this if you have any plan to fly with your cat, ever. By Petmate’s own listed dimensions, this kennel exceeds United’s under-seat maximum on every axis, and rigid plastic can’t compress to close that gap the way fabric can.
Reserve the SturdiBag XL for the Car, Not the Cabin
Maine Coon owners, this section is for you. The SturdiBag XL’s 40-pound rating is double the group average, per Sturdi Products’ own size chart, and the 600-denier fabric and multiple entry points make loading a large, opinionated cat manageable. At roughly $109 for the current Pro version — one retailer had it temporarily sold out across all colors at the time of this check, so confirm stock before counting on it — it’s also the cheapest carrier here per pound of rated capacity.
What it isn’t: a flying carrier. Sturdi Products’ own site states plainly that XL and XXL sizes are too large for in-cabin use, which the marketing for the SturdiBag line as a whole — built around flex-height cabin fit — doesn’t advertise loudly at this size.
Skip this if you bought it expecting to fly. Buy it if your cat is simply too big for anything else here and your trips are by car.
A carrier that fits at home can still fail at the gate if it doesn’t match the dimensions printed on your specific airline’s policy page.
Do the Price-Per-Pound Math Before You Compare Stickers
Sticker price alone mixes up two different things: what you’re paying, and how much cat that payment actually covers. Divide each carrier’s price by its rated weight capacity and the ranking moves.
Carrier
Price
Rated capacity
Price per rated pound
SturdiBag XL
~$109
40 lb
$2.73
Sherpa Original Deluxe
$69.99
22 lb
$3.18
Petmate Two Door Top Load
$64.99
15 lb
$4.33
Sleepypod Air
$214.99
18 lb
$11.94
The Sleepypod costs nearly 4.4 times as much per pound of cat as the SturdiBag. That premium buys the only independent crash-test rating in the group and the smartest under-seat engineering — worth it if you fly regularly, hard to justify for two vet visits a year.
Weigh the Axes Yourself If You Disagree
Four things decide this purchase: whether the carrier actually fits a cabin seat by the numbers above (35%), price per rated pound (25%), how easy the carrier is for a stressed vet visit (20%), and whether the maker backs it with anything beyond a standard return policy (20%). Score each carrier 0-100 on each axis and weight them, and here’s where they land.
Axis (weight)
Sherpa
Sleepypod
Petmate
SturdiBag XL
Verified cabin fit (35%)
70
100
10
0
Price per rated pound (25%)
86
23
63
100
Vet-visit usability (20%)
60
40
100
70
Named backing beyond a standard return (20%)
85
95
30
30
Weighted total
75.0
67.7
45.3
45.0
Disagree with 35% on cabin fit? Cut it to 15% and split the difference across the other three, and the SturdiBag XL jumps past both Petmate and Sleepypod on pure economics. That’s a defensible way to shop if flying was never the plan. It’s just not most cat owners’ plan, which is why the weight sits where it does.
The Sherpa wins on the tally above, and it’s the carrier most households should actually buy: it clears the airline math, it’s mid-priced, and it’s the only carrier with a named refund program behind it. The SturdiBag XL finishes last here not because it’s poorly made — it’s the cheapest per pound of cat by a wide margin — but because its own maker rules out the exact use case, in-cabin flying, that the marketing around the SturdiBag line leans on. Buy it for a big cat’s car rides, not for a flight. Round out the rest of the setup with our litter box comparison and our cat tree stability testing, or browse the full Cats category.
Some links on our site may earn us a commission at no extra cost to you. Learn more.
Updated September 2026: Every price, BTU rating, cooking area, and warranty term below is re-verified against each maker’s own current product page. Propane cost-per-hour uses the U.S. Energy Information Administration’s most recent published national average alongside current 20-pound exchange-tank pricing, since backyard grillers rarely buy propane the way the EIA measures it.
The direct answer: the cheapest grill here, Char-Broil’s $299 Performance Series 4-Burner, has more cooking area than the $499 Weber Spirit E-325. Price and capacity don’t move together in this lineup, and neither does price and BTU output. Three brands, four grills, one $299-to-$1,949 spread — the Ledger below scores what that money actually buys, not what the box promises.
Char-Broil Doesn’t Sell a “Performance 475” Anymore
“Performance 475” is a name that circulates widely in grill comparisons and forum threads — a 4-burner, 40,000-BTU model with a side burner. Char-Broil’s current site doesn’t carry that name or that spec sheet. The Performance Series 4-Burner lineup sold today runs a different number: 32,000 BTU on the base cart model (463351125, $299), climbing to 40,000 BTU only on the pricier FlavorMax cabinet variants with a side burner ($399). The 475 name is gone; the closest current match is a different machine at a different price.
Burner output tells you how fast the grate gets hot. It doesn’t tell you what an hour of that heat costs.
That kind of renaming is common enough in this category that it’s worth checking before you shop, not after: a model number or nickname that shows up in an old review can simply stop being sold. Char-Broil’s own site is the record here, and as of this update it lists the base cart grill at $299 with 435 square inches of primary cooking area (565 total with the warming rack) and a 1-year overall warranty.
The Ledger: Five Weighted Axes for Four Grills
Cooking area and running cost matter more than sticker price alone, so they carry real weight below — but warranty depth gets the most, because a grill’s cookbox is the part that actually determines how long you own it.
Axis (weight)
Weber Spirit E-325
Char-Broil Performance 4-Burner
Weber Genesis E-335
Napoleon Prestige 500 RSIB
Warranty depth (25) — years and parts actually covered, per each maker’s own warranty page
21/25 — “Lifetime” on hood, lid, grids and tube burners, but sear plates, infrared burners and “all other parts” drop to 15 yrs and electronics to 3 yrs, with corrosion and discoloration excluded outright
Cooking area per dollar (25) — total sq in ÷ price
13/25 — 450 sq in / $499 = 0.90 sq in per $1
25/25 — 565 sq in / $299 = 1.89 sq in per $1, the best in the field
10/25 — 513 sq in (primary; Weber’s own page lists no separate total) / $1,249 = 0.41 sq in per $1
7/25 — 900 sq in / $1,949 = 0.46 sq in per $1
BTU per dollar (20) — rated heat output ÷ price
15/20 — 31,000 BTU / $499 = 62 BTU per $1
20/20 — 32,000 BTU / $299 = 107 BTU per $1, cheapest heat in the lineup
Propane running cost (15) — cost per hour at full burn, exchange-tank pricing
13/15 — about $1.92/hr on the main burners
13/15 — about $1.98/hr
8/15 — about $2.41/hr on mains alone, $3.95/hr with the sear and side burners lit
5/15 — about $5.07/hr with the infrared side and rear burners running
Claim discipline (15) — does the published spec sheet match what a buyer can verify
13/15 — specs match Weber’s own current listing exactly
6/15 — the “475” name and its 40,000-BTU spec no longer match any current SKU
13/15 — specs match, though the site doesn’t publish a total cooking-area figure the way competitors do
12/15 — specs match, but “Lifetime Limited Warranty” as a headline undersells how much of the grill actually drops to 15 or 3 years
Total
72
72
63
53
Weber Spirit and Char-Broil’s Performance 4-Burner tie at 72. Disagree with a weight? The inputs are all above — cut warranty to 15 and hand the difference to cooking area, and Char-Broil pulls ahead outright. Raise warranty to 30 and Weber’s Genesis closes most of the gap. Re-run it with your own priorities.
Cooking Area Doesn’t Track Price the Way You’d Expect
Here’s the arithmetic in one place, main-burner BTU only (sear and side burners excluded, since they’re not always lit), against each maker’s own currently published spec.
Grill
Price
Total Cooking Area
Area per Dollar
Main Burners, BTU
BTU per Dollar, Mains Only
Char-Broil Performance 4-Burner
$299
565 sq in
1.89 sq in
32,000
107 BTU
Weber Spirit E-325
$499
450 sq in
0.90 sq in
31,000
62 BTU
Weber Genesis E-335
$1,249
513 sq in*
0.41 sq in
39,000
31 BTU
Napoleon Prestige 500 RSIB
$1,949
900 sq in
0.46 sq in
82,000†
42 BTU
*Weber’s own product page for the E-335 lists only a primary cooking area, not a combined total with the warming rack the way Char-Broil and Napoleon do. A similarly named “E-335s” variant sold by some retailers lists 787 square inches — that’s a different SKU with a larger sear zone and prep table, not the plain E-335 priced here. Compare model numbers, not just names. †Napoleon doesn’t break out a mains-only figure the way Weber and Char-Broil do; 82,000 BTU is the grill’s full rated output including the infrared side and rear burners, so its per-dollar figure isn’t strictly apples-to-apples with the other three’s mains-only numbers — if anything, that makes its BTU-per-dollar look better here than a mains-only figure would.
The $1,650 that separates the cheapest grill from the most expensive buys 50,000 more BTU of rated heat and 335 more square inches of cooking area between the extremes — but not proportionally, and not on the axis most buyers assume. Napoleon’s price climbs mostly on burner count, infrared technology, and stainless construction, not on cooking area or heat-per-dollar. If capacity and running cost are what you’re buying, the Char-Broil and the Weber Spirit are doing more with less money than the two grills above them.
What “Lifetime” Actually Covers on the Napoleon
Napoleon markets the Prestige 500 RSIB with a “Lifetime Limited Warranty,” and on paper it’s the strongest single word in this comparison. Read the warranty document itself and the picture is more layered. Lifetime coverage applies to the hood, stainless lid, porcelain-enamel lid, aluminum castings, stainless cooking grids, and stainless tube burners. Sear plates, the infrared rotisserie burner, and the infrared side/rear burners drop to 15 years, as does the catch-all “all other parts.” Electrical components get 3 years — the same ceiling Weber puts on its own electronics. And the warranty explicitly excludes “scratches, dents, corrosion, deterioration of painted and plated finishes, discoloration caused by heat” and UV exposure. None of that makes “Lifetime” false. It does mean the headline word covers roughly half the grill, and the other half runs on ordinary consumer-appliance terms.
Weber doesn’t use the word “lifetime” at all, and its own numbers turn out to be close: 12 years on the Genesis cookbox and lid, 10 years on burner tubes, grates, and Flavorizer bars, 5 years on whatever’s left, 3 years on electronics. Framed as a headline claim, Napoleon wins. Framed as a components table, the two are within a few years of each other almost everywhere that matters.
The Tank You’re Actually Buying Propane From
The propane math above assumes exchange-tank pricing, not the EIA’s residential bulk-delivery rate most propane cost calculators quote. That distinction matters: most backyard grillers refill by swapping a 20-pound tank at a hardware store or gas station for $17 to $25, and exchange programs commonly fill those tanks to about 15 pounds, not the full 20 — a detail printed in the fine print of the exchange terms, not on the cage the tanks sit in. Fifteen pounds of propane is roughly 3.5 gallons; at $20 a swap, that works out to about $5.65 a gallon, more than double the EIA’s most recent published national residential average of $2.67 (week ending March 30, 2026, the agency’s last release before its seasonal pause). Buying a refill tank instead of swapping, where available, is usually the cheaper way to get a full 20 pounds. The hourly costs in the Ledger use the $5.65 exchange-tank rate, since it’s what most of this comparison’s readers are actually paying at the register.
Skip Each of These If
Skip the Weber Spirit if cooking area matters more to you than the Weber badge — the Char-Broil beats it on space and BTU-per-dollar at $200 less. Skip the Char-Broil if you want a grill you’re not thinking about replacing burner-by-burner in year two; a 1-year overall warranty and 1-year grates are the shortest terms here by a wide margin. Skip the Weber Genesis if raw capacity per dollar is the deciding factor — its 513 square inches cost more per square inch than any grill in this comparison except the Napoleon. Skip the Napoleon if $1,949 is simply more than this decision is worth to you, or if “Lifetime Limited Warranty” was doing more work in your head than it does on the actual coverage sheet.
The Tie-Breaker: Char-Broil, Unless the Cookbox Matters More Than the Price
Weber Spirit and Char-Broil tied the Ledger at 72, and a tie this close comes down to what you value that the tally can’t fully weigh. If this is Nate’s money: the Char-Broil. More cooking area, cheaper heat, and a $200 saving that buys most of a season’s worth of propane on top. The 1-year warranty is real and it’s the honest tradeoff for that price — but a grill used a few dozen times a year and covered, kept dry, and not abused tends to outlive a 1-year clock anyway. Where that calculus flips: if you grill often enough that burner and grate longevity actually gets tested, Weber’s 10-year burner-tube coverage on the Spirit, or Genesis if the budget stretches, is worth the premium. And if 900 square inches of cooking space for a crowd is the actual requirement, no amount of Ledger math changes that only the Napoleon delivers it.
Prices are approximate, current as of publication, and change frequently — confirm the latest price with the retailer before buying.
Some links on our site may earn us a commission at no extra cost to you. Learn more.
Updated August 2026: each model’s certified annual kilowatt-hours pulled from the Energy Star product database, running 270 to 323. Capacities, prices and lab results re-verified at the same time. Added the Galanz retro-refrigerator recall of 30 July 2026 and the Curtis International recall covering Frigidaire-branded minifridges, neither of which names a model on this page.
Buy the Midea WHD-113FSS1 at $183.99 if you can find one in stock, and the Magic Chef HMCR320WE at $269.99 if you cannot. Those are the only two of these four fridges that will tell you how much fresh-food space is actually inside them.
The Frigidaire EFR840 is the loser here, and not for the reason you would guess. It is the second-cheapest fridge on the list. It is also not made by Frigidaire, does not publish a capacity split, and does not appear in the federal energy database at all.
Four years of electricity: $198 to $237, from the labels themselves
A dorm fridge runs 8,760 hours a year and nobody unplugs it for reading week. The annual figures below are not estimates. They are each model’s certified consumption in the Energy Star certified-product database, searched by model number, alongside the federal baseline each one had to beat.
Model
Price today
Certified kWh/year
Federal baseline
Electricity, 4 years
Four-year total
Midea WHD-113FSS1
$183.99
270 (certified 2014)
358
$198.07
$382.06
Magic Chef HMCR320WE
$269.99
323 (certified 2018)
359
$236.95
$506.94
Galanz GLR31TBEER
$289.99
320 (certified 2018)
358
$234.75
$524.74
Frigidaire EFR840
$198.00
not in the database
—
cannot be computed
cannot be computed
The arithmetic uses 18.34 cents per kilowatt-hour, which the Energy Information Administration reported as the June 2026 US residential average. So 270 times $0.1834 is $49.52 a year and $198.07 over four; 323 gives $59.24 and $236.95; 320 gives $58.69 and $234.75. Add the purchase price and you get the last column.
Read the Midea row twice. Four years of electricity costs $198.07. The fridge costs $183.99. You will spend more powering it than buying it, and that is the model with the lowest consumption of the three that are listed.
Then read the Frigidaire row. Searching Energy Star for EFR840 and for EFMIS840 returns nothing. There is no certified annual figure to plug in, so the cheapest-looking four-year total on this page is the one that cannot be calculated. On a machine that runs continuously for four years, that is not a footnote.
Prices are approximate, current as of publication, and change frequently — confirm the latest price with the retailer before buying.
Electrolux does not build the Frigidaire, and three other things the boxes get wrong
The claim
What the primary source says
Verdict
"Frigidaire" EFR840
The manufacturer is Curtis International Corp, which licenses the Frigidaire name from Electrolux. Curtis’s own product page carries the licence notice in its footer.
The badge is a licence. Warranty claims, recall notices and support all go to Curtis, and you will not find them by searching Frigidaire.
Galanz: 3.1 cu ft. Frigidaire: 3.2 cu ft.
Neither maker publishes a fresh-food and freezer split anywhere I could find. Midea publishes 2.2 plus 0.9. Magic Chef publishes 2.24 plus 0.93.
Two of the four will sell you a number for the whole box and decline to say how much of it holds your milk.
Frigidaire EFR840: "3.2 CU. FT. CAPACITY"
The same Curtis product page also states 2.8 cu ft total net capacity, without reconciling the two.
The maker’s own page gives two different sizes for one fridge. Believe the smaller one.
Magic Chef: 3.2 cu ft total
2.24 plus 0.93 is 3.17
A rounding gap, not a lie. Noted because it is the only one of the four where the parts and the total can be checked against each other at all.
Roundups calling the EFR840’s freezer the standout
Reviewed’s award for the EFR840 is Best Value Retro Mini Fridge, a value commendation. Its measured note is that the fridge held consistent temperatures with some wobbling over time. No freezer figure is published for it.
Wrong superlative. On Reviewed’s own numbers the tightest freezer belongs to the Magic Chef.
Two of these four sit inside an open recall on their maker’s other models
No model on this page is recalled. Two of the four brands have live recalls on adjacent minifridges, and both are serious enough that you should know the model numbers before you buy a used one or accept a hand-me-down.
On 30 July 2026, CPSC announced that Galanz Americas recalled about 121,680 retro refrigerators over an internal short circuit that can ignite. Thirty-four fires and one death were reported. The recalled units are the 7.4 and 7.6 cubic-foot models — BCD-215V-62H, GLR76TRDER, GLR76TBKER, GLR76TBEER and GLR76TWEER — with date codes from December 2018 to December 2020. The 3.1 cubic-foot GLR31 series on this page is not among them.
Curtis International, the actual maker of the Frigidaire-branded fridges, recalled 634,000 minifridges in August 2025 over smoking, sparking and fire, with two smoke-inhalation injuries and more than $700,000 in property damage, then expanded it on 15 January 2026 by roughly 330,000 more units. Recalled models are EFMIS121, EFMIS129, EFMIS137, EFMIS149 and EFMIS175. The EFR840 is not named in either action.
Midea’s large recall is dehumidifiers, announced in 2016, and has nothing to do with refrigeration. A search returned no compact-refrigerator recall for Magic Chef.
Five degrees against a federal standard of zero
The FDA and USDA both give the same two numbers: refrigerator at or below 40°F, freezer at 0°F. Those are the lines a compartment either holds or does not.
Reviewed, in a roundup updated 13 May 2026, loads each fridge with water ballast and logs temperature consistency in both compartments across a week. Against that method:
Model
What Reviewed measured
Against the standard
Magic Chef HMCR320WE
Fridge within a degree or two of target; freezer within a single degree
The only one here described as holding its freezer tight
Frigidaire EFR840
Consistent fridge temperatures, some wobbling over time. No freezer figure given.
Fridge acceptable. Freezer unmeasured in what is published.
Midea WHD-113FSS1
Default settings ran warm in the fridge compartment
Ships mis-set. Fixable with the dial and a $6 thermometer, but you have to do it.
Galanz GLR31TBEER
Freezer averaged 5°F instead of 0°F and ranged 2°F to 8°F — the least consistent freezer in the group
Fine for ice packs. Not a freezer for a month of meal prep.
Worth knowing what "compact" even means here: the Department of Energy defines the class in 10 CFR 430.2 as any refrigerator or refrigerator-freezer under 7.75 cubic feet. All four of these sit at less than half that, which is why their freezer compartments behave nothing like the one at home.
Every measurement above belongs to Reviewed, named in the sentence. Nothing on this page came off a bench of ours, and where a figure does not exist the table says so instead of filling the cell. Our editorial standards explain the sourcing rules.
Weighting measured cold against four-year cost
Five axes, weights chosen for a student who will own one fridge for four years and cannot replace it mid-semester. Scores out of 10.
Axis (weight, and why)
Magic Chef HMCR320WE
Midea WHD-113FSS1
Galanz GLR31TBEER
Frigidaire EFR840
Measured temperature control, 25%. A fridge has one job and it does it unattended.
10 — within a single degree in the freezer
5 — ships running warm
3 — freezer at 5°F, ranging 2°F to 8°F
6 — fridge consistent, freezer unmeasured
Four-year cost, 25%. Purchase plus the power it draws for 35,000 hours.
5 — $506.94
9 — $382.06
4 — $524.74
5 — cheap to buy, uncomputable to run
Fresh-food space you can verify, 25%. You buy a fridge for the space inside it.
9 — 2.24 cu ft published
9 — 2.2 cu ft published
3 — total only
2 — total only, and stated two ways
Published energy accountability, 15%. A certified figure is a promise someone can be held to.
8 — 323 kWh against a 359 baseline
10 — 270 kWh, lowest here
8 — 320 kWh against a 358 baseline
1 — absent from the database
Brand accountability, 10%. Who answers the phone, and what their other models are doing.
7 — no recall found
5 — no fridge recall; the certification dates to 2014
4 — one-year warranty excluding commercial use; sibling retro line under recall
3 — licensed badge, and the actual maker has ~964,000 minifridges recalled
Weighted tally
7.9
7.8
4.1
3.7
Magic Chef’s 7.9 comes from 2.50 on cold, 1.25 on cost, 2.25 on verifiable space, 1.20 on energy and 0.70 on brand. Midea’s 7.8 is 1.25, 2.25, 2.25, 1.50 and 0.50.
Move four-year cost up to 40 percent and drop verifiable space to 10, which is how you should score it if the fridge is going in a single room and the money is tight, and Midea takes it clearly: 7.8 against 7.3. Nothing moves the bottom two. The Galanz and the Frigidaire lose on different failures and both losses are structural.
Three cubic feet is the whole box. How much of it holds food is a number two of these four makers will not print.
The Magic Chef holds the tightest freezer of the four
Reviewed’s note on the HMCR320WE is the strongest measured claim in this whole comparison: fridge within a degree or two of target, freezer within a single degree. For a box that guards a week of groceries while you are in class, that is the trait worth paying for.
It also publishes its capacity honestly at 2.24 plus 0.93, it is Energy Star certified at 323 kWh a year, and MC Appliance Corp has no compact-refrigerator recall on the record. At $269.99 it is a Home Depot exclusive, which is worth knowing if you were planning to price-shop it.
The recurring complaint in its 1,839 Home Depot reviews, which average 4.5 stars, is a door that needs a firm push to latch. I could not get a count of how many reviews say that against how many do not, so treat it as a known gripe rather than a measured failure rate.
Skip the Magic Chef if you are splitting a fridge with a roommate. Two people’s groceries do not fit in 2.24 cubic feet and no amount of temperature precision fixes that.
The Midea is the cheapest to own and the hardest to find
Two hundred and seventy kilowatt-hours a year is the lowest figure on this page by 50, and $382.06 over four years is $125 less than anything else here. It publishes a real capacity split. It is the value pick on arithmetic and it is not close.
Now the problem. Best Buy lists it at $183.99 and flags it as no longer available in new condition. Midea’s own store page for the WHD-113FSS1 loads without a price or a checkout. The model is still listed by its maker, so this is not a discontinuation, but it is not a fridge you can reliably order in August either. Check stock before you plan around it.
One more caveat: its Energy Star certification dates to 2014. A twelve-year-old certification on a model still being sold is not a defect, but it does mean you are buying a design from a different decade.
Skip the Midea if you need the fridge on move-in day and cannot wait on stock, or if you will not spend ten minutes turning the thermostat down and checking it with a thermometer.
The Galanz has the warmest freezer Reviewed measured
Five degrees average, swinging between 2 and 8, against a federal standard of zero. The separate freezer door is real and the retro shell is genuinely nice, but the compartment behind that door is an ice-pack cabinet, not a freezer.
At $289.99 it is also the most expensive fridge here and, at $524.74 over four years, the most expensive to own. Galanz publishes no capacity split, and its one-year warranty excludes commercial use — language worth reading if the fridge is going into a shared lounge rather than your own room.
Skip the Galanz if anything you buy frozen needs to still be frozen next month.
The Frigidaire is the cheapest to buy and the only unknown to run
At $198.00 with 540 Home Depot reviews averaging 4.1 stars, the EFR840 looks like the sensible budget answer. Three things argue against it, and they compound.
It is absent from the Energy Star database, so nobody can tell you what four years of it costs. Its maker’s own page states its capacity as both 3.2 and 2.8 cubic feet and never splits fridge from freezer. And the company that actually builds it, Curtis International, has roughly 964,000 minifridges under recall across two 2026 actions on other model numbers.
None of that is proof this fridge is bad. All of it is a refusal to publish the numbers the other three publish, and on a purchase you keep for four years that refusal is the finding.
Skip the Frigidaire if you want to know what you are agreeing to before you agree to it.
Where the top two tie, and how I break it
Magic Chef 7.9, Midea 7.8. A tenth of a point is noise and I am not going to pretend it is a verdict.
My money goes to the Midea WHD-113FSS1, on the $125. Four-year ownership is where a dorm fridge decision actually lives, and the Midea wins it by more than the price of the fridge you would have bought instead. It also carries the only energy figure on this page that is both the lowest and independently checkable.
That comes with a condition, and it is a real one. If you cannot find a WHD-113FSS1 in stock — and in August you may not — buy the Magic Chef HMCR320WE and do not feel bad about it. It is the only fridge here whose freezer a lab has measured holding within a degree, and that is worth $125 to anyone who buys frozen food rather than ice.
Measure the loft clearance before the fridge arrives, not after it is standing in the hallway.
What dorm buyers ask before move-in
How much does a mini fridge cost to run for a year?
Between $49.52 and $59.24 for the three models here with certified figures, at the 18.34 cents per kilowatt-hour the Energy Information Administration reported for June 2026. That is 270 to 323 kilowatt-hours a year. Any source quoting under 200 for a 3-cubic-foot two-door model is estimating by size class rather than reading the label.
Will a mini fridge freezer keep ice cream frozen?
Only if it holds 0°F, which is the FDA and USDA figure for frozen storage. Of the four here, the Magic Chef is the one Reviewed measured holding its freezer within a single degree. The Galanz averaged 5°F and swung to 8°F, which will keep ice cream soft.
Who actually makes Frigidaire mini fridges?
Curtis International Corp, under a licence for the Frigidaire name from Electrolux. This matters practically: warranty service, support and recall notices are Curtis’s, and searching for Frigidaire recalls will not surface them.
Are any dorm mini fridges under recall right now?
Yes, though none of the four compared here. Galanz recalled about 121,680 retro refrigerators in July 2026 after 34 fires and one death, covering the 7.4 and 7.6 cubic-foot GLR76 and BCD-215V-62H models. Curtis International has about 964,000 Frigidaire-branded minifridges recalled across August 2025 and January 2026, covering the EFMIS121, 129, 137, 149 and 175. Check the model number on the back before you accept a used one.
Do this before you order
Search your exact model number in the Energy Star product finder and in the CPSC recall database. Both are free, both take under a minute, and between them they answered more about these four fridges than every product page did.
Some links on our site may earn us a commission at no extra cost to you. Learn more.
Updated August 2026: TechGearLab’s measured average and peak temperatures are shown side by side for each oven — peaks run near 1,100°F on three of the four. Ooni’s Koda 2 is sold out in every colour on Ooni’s own store. The Solo Stove Pi Prime is $359.99 on Solo Stove’s own site. No cost-per-pizza-night figure appears here: three of these four makers publish no gas consumption rate to compute one from.
Weighing Ooni vs Gozney vs Solo Stove, the Solo Stove Pi Prime is the pizza oven to buy. It scores 85.5 out of 100 on the five weighted axes below, eight points clear of anything else here. That is because it pairs strong heat with the lowest price and the lightest body at 30 pounds. The Gozney Arc runs hotter than any other outdoor pizza oven on this page and finishes last on the tally at 75.8, which is the whole argument of this comparison. Heat is a quarter of the score. It is not the score.
Cooking outside in Florida heat, I’d rather deal with something light I can move and store easily. A few extra degrees of peak heat don’t matter much if the oven that hits it costs more and weighs more.
The pick: Solo Stove Pi Prime, $359.99. Cheapest, lightest, 911°F average. Twelve-inch stone is the catch.
If you want 14-inch pies: Gozney Arc, $799.99. The hottest oven measured, and 47.5 pounds of it.
If you want fire: Ooni Karu 2, $449. Wood, charcoal, or a gas burner Ooni sells separately and does not price on the product page.
The one you cannot have: Ooni Koda 2, $499, sold out in all three colours on ooni.com.
Ooni vs Gozney vs Solo Stove: which pizza oven wins the tally?
After hurricane seasons of hauling gear inside before a storm, the Solo Stove Pi Prime is the one I’d point a friend to, because 30 pounds is easy to carry alone. The Solo Stove Pi Prime wins this five-axis tally with a weighted score of 85.5 out of 100, eight points ahead of anything else in this comparison. It pairs the third-hottest average temperature in the group with the lowest price and lightest body of the four ovens.
Every axis below is scored 0 to 10, multiplied by its weight, and summed. Price is scored by dividing the cheapest oven’s price by each oven’s price and multiplying by ten. So the $359.99 Pi Prime scores 10.0 and the $799.99 Arc scores 4.5 ($359.99 ÷ $799.99 × 10). Temperature is scaled the same way against the hottest average. Disagree with a weight? Change it and re-run the row. That is what the numbers are for.
Axis (weight)
Gozney Arc
Solo Stove Pi Prime
Ooni Koda 2
Ooni Karu 2
Measured average temperature (25%)
10.0
9.4
9.6
8.0
Pizza size (20%)
10.0
6.5
10.0
6.5
Price at the maker (25%)
4.5
10.0
7.2
8.0
Availability and warranty terms (15%)
8.0
8.0
3.0
6.5
Weight and fuel flexibility (15%)
5.0
8.0
7.0
9.5
Weighted total
75.8
85.5
77.0
77.0
Why Do Temperature and Price Each Carry 25 Percent?
Temperature and price each carry 25 percent because they are the two things that actually separate these machines. Everything here clears the 750°F a decent pizza needs, and the price spread is more than two to one. Pizza size takes 20 percent because a 12-inch stone is a hard ceiling, not a preference, if you are feeding four people in one bake. Availability and warranty take 15 percent, which is more than most comparisons give them and is the reason the Koda 2 does not win this page. Weight and fuel flexibility take the last 15 percent.
The Arc wins two axes outright and still finishes last. It costs $440 more than the Pi Prime and gives back two inches of stone and a slightly hotter fire. That is a real thing to want. It is not a 2.2x thing to want.
Do Ooni, Gozney, and Solo Stove Ovens Hit Their Advertised Heat?
Gozney, Ooni and Solo Stove all advertise the same ceiling: 950°F. All four ovens on this page carry it. When one company and then a second and then a third arrive at the identical round number, that number is a specification of the marketing, not of the oven.
TechGearLab is the only outlet I could find that publishes discrete measured air and stone temperatures for these specific models, in its pizza oven rankings last updated November 2025. Its figures do not line up with the claim in either direction, and the gap runs both ways.
Oven
Maker’s stated max
Measured average air
Measured peak air
Measured average stone
Gozney Arc
950°F
969°F
1,105°F
871°F
Ooni Koda 2
950°F
935°F
1,102°F
848°F
Solo Stove Pi Prime
“in excess of 950°F”
911°F
1,100°F
841°F
Ooni Karu 2
950°F
776°F
not published
825°F
Three readings out of that table. The peaks beat the claim by 150 degrees or more, so the makers are underselling the top of the range. The averages sit at or below the claim, which is the number that matters, because you cook at the average and not at the spike. And the Karu 2 is the only oven here whose stone runs hotter than its air, 825°F against 776°F. That is not a transcription error. It is what a wood fire under a stone does.
How Reliable Is the Ooni, Gozney, and Solo Stove Heat Data?
These are comparative in-house measurements under an undisclosed protocol, from a single outlet, and that is one honest limit worth flagging. TechGearLab describes using an infrared thermometer for the stone readings. It does not publish how the air averages were taken, over what interval, or at what burner setting. Anyone telling you a consensus of laboratories has ranked these ovens is counting to one.
What Do Ooni, Gozney, and Solo Stove Ovens Cost at the Maker?
The Ooni Koda 2 lists at $499 but is sold out in every colour on Ooni’s own store. The Solo Stove Pi Prime remains the cheapest of the four at $359.99. Two of them have moved since this page was first written, in opposite directions, and one oven cannot be bought from the company that makes it.
Oven
Price at the maker
Stock
Max pizza
Weight
Warranty
Gozney Arc
$799.99
Available
14"
47.5 lb
1 year, 5 with registration
Solo Stove Pi Prime
$359.99 (from $399.99)
Available
12"
30 lb
Lifetime claimed; exclusions not published
Ooni Koda 2
$499.00
Sold out, all three colours
14"
35.3 lb
1 year, 5 with registration
Ooni Karu 2
$449.00
Available
12"
33.7 lb
1 year, 5 with registration
Prices are approximate, current as of publication, and change frequently — confirm the latest price with the retailer before buying.
Read the warranty column twice. Ooni’s five years is conditional on registering within 60 days and on having bought the oven from Ooni directly. Buy a Karu 2 at Costco and you have a one-year warranty, whatever the marketing implies. Solo Stove says lifetime, but its warranty terms page returned a 404 when I went looking for the exclusions. So I cannot tell you what that lifetime excludes, and neither, from that page, can you.
What I could not price at all: gas. Gozney publishes a consumption rate for the Arc, 4.5 kW or 330 grams an hour, in its help centre. Ooni publishes no consumption figure for the Koda 2 or either Karu, and Solo Stove publishes none for the Pi Prime. So a cost-per-pizza fuel number for three of these four ovens would be a number I made up, and it is not in this comparison.
Gozney Arc: the highest ceiling, the lowest total
The Arc took TechGearLab’s Editors’ Choice and posted the highest average of the eight ovens in that test, 969°F, with a peak of 1,105°F. Fourteen-inch stone, 20 mm thick and removable. It is the only oven here I would call furniture rather than equipment.
It is also $799.99, up from the $699.99 Tom’s Guide recorded in September 2024. Prices in this category are moving up, not down, which is worth knowing if you have been waiting for a sale. At 47.5 pounds it lives where you put it.
Skip this if you have not yet established that you will fire an outdoor oven more than a handful of times a year. Eight hundred dollars is an expensive way to discover you are a four-pizzas-a-summer household.
Solo Stove Pi Prime: lightest, cheapest, and the winner here
The Solo Stove Pi Prime is the cheapest and lightest oven in this comparison at $359.99 and 30 pounds, with a measured average temperature of 911°F, the third-hottest of the four ovens tested here.
911°F average, 1,100°F peak, 30 pounds, $359.99. It gives up 58 degrees of average heat to the Arc and takes $440 back. TechGearLab gave it the Best Buy award. Its own optimal cooking range for the oven, 750°F to 850°F, sits below the average the thing actually holds.
The stone is 12 inches. That is one pizza for one adult, and it is the reason this oven does not win every axis. Gas only, no wood path. Solo Stove sells a separate wood-burning Pi Prime, which is a different product rather than a conversion kit.
Skip this if you want one shared 14-inch pie per bake rather than a queue of personal ones. Cooking in rounds is fine for two people and tiresome for five.
Ooni Koda 2: fourteen inches, no award, no stock
The Koda 2 posts the second-highest average in the field at 935°F and peaks at 1,102°F, on a 14-inch stone, for $499. On the numbers it is the cheapest route to a large pie. TechGearLab still ranked it sixth of eight and gave it nothing. Its reviewer’s complaint was the control: a dial with no temperature markings, no recommended settings, no indication of where the useful zone is.
Then there is the part that decides its score here. Every colourway on Ooni’s own product page reads Sold Out with a notify-me button. It is not discontinued, and it still sits at $499 in Ooni’s pizza oven collection alongside the Koda 2 Pro and Koda 2 Max. You simply cannot buy it from Ooni today. Ignore the scarcity counter on that page, which reads “only 2,428 left” on an item every variant of which is out of stock.
Skip this if you need the oven this month. Watch the stock page instead of paying a reseller a premium for the same box.
Tending a pizza by hand inside a lit oven.
Ooni Karu 2: wood, charcoal, or a burner Ooni won’t price
The Ooni Karu 2 is the only oven on this page that burns wood, and it took TechGearLab's Top Pick for multi-fuel. It is 33.7 pounds and $449, and its 776°F average air temperature is the lowest here by 135 degrees while its stone runs the second-warmest of the four.
The gas burner is sold separately and Ooni does not publish a standalone price for it on the product page. I can infer roughly $120 from the gap between the bundle and the bare oven, but inferring a price from a bundle delta is not the same as knowing one. So treat the Karu 2 as a $449 oven that costs more than $449 the moment you want it to behave like the other three.
Skip this if you want to cook on a weeknight. Wood wants tending between pizzas, and the 12-inch stone means you will be tending it more than once.
Have Ooni, Gozney, or Solo Stove ovens been recalled?
I searched CPSC recall notices for Ooni, Gozney and Solo Stove and found nothing for any of them. That is “no recall found,” not “no recall exists”. The CPSC search endpoint refuses automated queries, so this was targeted searching rather than an exhaustive sweep of the database.
What did turn up is adjacent and recent. On 9 July 2026, Conair recalled about 12,660 Cuisinart Propel+ four-burner grills, model CGG-6331. The tempered glass window in the grill’s pizza oven can shatter during use. Thirty-seven reports of shattered glass, one fire, no injuries reported, $500 refund.
No oven on this page has a glass door. Ooni’s Karu 2 Pro, which is the next model up from the Karu 2, does have a glass door, in borosilicate rather than tempered glass. It has no recall and no reported failure pattern. If you are shopping one rung above this comparison, that is the honest answer rather than silence.
Ooni Koda 2 vs Ooni Karu 2: which should you buy?
The Ooni Koda 2 and the Ooni Karu 2 finish level at 77.0. Not close. Level. They get there by opposite routes. The Koda 2 buys heat and stone size and loses on stock, while the Karu 2 buys fuel flexibility and portability and loses on temperature and pie size.
My money goes to the Karu 2, on one condition: that you actually intend to burn wood in it. Wood is the only thing on this page a $359.99 Pi Prime cannot do. It is the only reason to accept a 12-inch stone and the lowest average temperature of the four. Budget above the $449, because the gas burner is extra and unpriced.
If you want gas, I would not buy either of these two. Buy the Pi Prime the tally already picked.
Is the Gozney Arc Lite worth considering instead?
Gozney now sells the Arc Lite, a 12-inch gas oven, at $349.99 down from $399.99. It undercuts the Pi Prime by ten dollars, from the company that builds the hottest oven in the field.
It is not in the tally because nobody has published measured temperatures for it. Scoring an oven on a maker’s own claim, after spending this page explaining that the claim is a round marketing number, would be dishonest. So: a live candidate, no evidence yet, and you should know it exists before you spend $359.99 on the winner. Gozney’s Tread at $499.99 and the Roccbox at $399.99, now well below the $500 TechGearLab still lists, are in the same position.
What Should You Check Before an Ooni, Gozney, or Solo Stove Buy?
Prices in this category move, and the cheapest 14-inch oven in the group went out of stock at its own maker. Open the maker’s page, confirm the oven is buyable and confirm the number, then come back and re-run the tally with today’s figures. The weights are yours to change.
Some links on our site may earn us a commission at no extra cost to you. Learn more.
Feeding decisions are personal and medical circumstances vary. This comparison covers equipment only. For latch issues, supply concerns, pain, or any feeding question, talk with a lactation consultant or your healthcare provider. And follow each pump manufacturer’s instructions for use and cleaning.
Updated August 2026: Elvie replaced the Stride with the Stride 2 and roughly doubled its suction claim on the box. Momcozy replaced the S9 with the S9 Pro and stopped publishing a suction figure at all. The Medela unit with independent lab data has become an insurance-channel purchase rather than a general-retail one. Every price, model number and score below reflects what you can actually buy today.
The Spectra S1 Plus is the only breast pump whose independently measured numbers still describe the exact unit on sale. Those numbers are 250 mmHg of suction and 0.55 ounces a minute. BabyGearLab measured them in testing last updated September 9, 2025, at $226. Momcozy’s S9 Pro, Elvie’s Stride 2 and the Medela you’ll actually find at a general retailer have all changed. That happened since any lab weighed in. This comparison is about what that gap is worth to you.
Prices are approximate, current as of publication, and change frequently — confirm the latest price with the retailer before buying.
Why Doesn’t a Breast Pump’s Lab Data Match What’s on the Shelf?
Because three of these four product lines have moved since anyone with a suction gauge checked them. BabyGearLab’s breast pump testing was last updated September 9, 2025. It’s the only instrumented, published measurement of suction and output I could find for this category. Two of the machines it measured have since been replaced by name, and a third has effectively changed sales channel. Here is each one, tested against what’s actually for sale.
None published — the S9 Pro’s spec sheet lists “9 levels” and battery figures, no mmHg or oz/min
Predecessor’s number no longer applies to what’s sold; successor publishes nothing to replace it
Medela Pump In Style with MaxFlow, Handsfree (general retail)
Medela Pump In Style with MaxFlow, corded — 160 mmHg, 0.24 oz/min
None published for the Handsfree variant
A differently built product; untested by any lab I could find
Read that middle column carefully. It isn’t that these pumps got worse, or that anyone lied. It’s that “tested” has an expiration date nobody prints on the box. Three names on this page have quietly outlived their own lab data.
The numbers behind a pump matter most months in, long after the box and its claims are gone.
Is the Elvie Stride 2’s claimed suction real?
Elvie’s own product page for the Stride 2 advertises “high maximum pressures from -270mmHg up to -300 mmHg”. That’s across 10 stimulation and 10 expression settings. The Stride it replaced — the one BabyGearLab actually put a gauge on — measured 170 mmHg. If the new number holds up, that’s not a refinement, it’s close to double. I could not find a published independent measurement of the Stride 2 itself. Not from BabyGearLab, and not from anyone else, as of this writing.
What Elvie does publish elsewhere checks out on its own terms: roughly three hours of pumping per charge. It also lists a 5-ounce capacity per cup and six flange sizes in the box. It claims a fit for 99 percent of nipple sizes. None of that is suction, and suction is the number this whole comparison turns on. At $269.99 list — commonly discounted, $202.49 at the time I checked — it’s the most expensive pump here. It’s also the only one whose central performance claim is entirely maker-reported.
Skip this if the claimed suction jump is the reason you’re buying. Nobody outside Elvie has confirmed it yet. You’d be paying the top price on this page for a number with one source.
Where did the Momcozy S9 Pro’s suction number go?
The original S9 has a real lab result: 230 mmHg of suction and 0.20 ounces a minute of output. BabyGearLab also gave it a “Loud” noise rating. The S9 Pro that replaced it publishes none of that. Momcozy’s own spec sheet for the Pro lists two modes and nine intensity levels. It also lists an 8-to-9-session battery life of roughly 270 minutes total. It lists a 30-minute auto shut-off and a noise claim under 45 decibels too. Every one of those is a maker figure, none of them suction or output.
That’s not necessarily a downgrade. It’s possible the Pro pumps harder than the 230 mmHg unit it replaced. It’s also possible it doesn’t. Dropping the number from the spec sheet is a convenient way to avoid finding out in public. Either way, at $119.99 the S9 Pro is the cheapest pump in this comparison by a wide margin. Its battery life — roughly 4.5 hours per charge — beats the Spectra’s and the Elvie’s published 3 hours each.
Skip this if you need to know your suction setting will still work at the flange size. And it needs to work at the pressure your last pump was calibrated to. There’s no published number here to calibrate against.
Why is the Medela BabyGearLab tested now hard to buy at retail?
This is the finding I didn’t expect. BabyGearLab measured the exact corded Pump In Style with MaxFlow at 160 mmHg and 0.24 oz/min. Those are the weakest numbers on that lab’s chart. That exact unit shows as no longer available new at a major electronics retailer I checked. Its remaining listings sit mostly with insurance-focused durable-medical-equipment suppliers rather than ordinary retail shelves, at a list price near $260.
Walk into general retail and the Medela you’ll actually find under the “Pump In Style” name is the Handsfree version. It’s currently listed around $130.99 — roughly half the price and a wearable-adjacent design. I could not find independent suction or output data for it anywhere. It isn’t the machine that got tested. Medela’s own current lineup has also split into Pro, Pro+, and Handsfree variants under the same “Pump In Style” name. That makes it easy to think you’re buying the reviewed pump when you’re buying a sibling product instead.
Skip this if you’re shopping general retail and assuming “Pump In Style” means the pump the reviews describe. Check the exact model name against your insurance supplier’s list before you decide it’s the same machine.
How Do Four Breast Pumps Score on What’s Verifiable?
The Spectra S1 Plus scores highest on this weighted ledger at 81.6. That’s largely because it’s the only pump whose lab-verified numbers still describe the unit sold today. The Medela Pump In Style Handsfree scores lowest at 37.9.
Four axes. Whether independent lab data actually describes the model on sale carries the most weight, at 40 percent. That’s because it’s the entire premise of trusting a spec sheet. Price is 25 percent. Published battery life per charge is 20 percent. That’s since a pump that dies mid-session doesn’t get scored on suction at all. And whether you can buy the exact reviewed unit at ordinary retail today, not through an insurance-only channel. That gets the last 15 percent.
How Is Each Axis Weighted?
Axis (weight)
Spectra S1 Plus $226
Momcozy S9 Pro $119.99
Elvie Stride 2 $269.99
Medela Pump In Style (Handsfree) $130.99
Verified current-model data (40%)
100 — exact model, Sept. 2025 lab data
20 — predecessor’s number doesn’t carry over; successor publishes none
15 — the tested unit is a different, harder-to-buy product
Price, cheapest = 100 (25%)
53.1
100
44.4
91.6
Published battery life per charge (20%)
66.7 — ~3 hrs
100 — ~4.5 hrs across 8–9 sessions
63.0 — ~3 hrs
15 — not published anywhere I could verify
Buyable as the reviewed unit at general retail (15%)
100
100
100
40 — the tested corded unit is now largely an insurance-channel item
Weighted total
81.6
68.0
46.7
37.9
Disagree with 40 percent on verifiability? Drop it to 15 and split the difference across price and battery. Momcozy then pulls ahead of Spectra on cost alone. That’s a defensible way to shop if you’ve decided a lab result from a discontinued model is better than none. I don’t think it is, which is why the weight sits where it does.
How Much Does Each Breast Pump Cost per Minute of Output?
The Spectra S1 Plus is the only pump here with a verified cost per speed. Its $226 price for 0.55 oz/min works out to $41 per 0.1 ounce-per-minute. That’s a figure nobody can compute yet for the Momcozy S9 Pro, Elvie Stride 2, or Medela Handsfree. None of those publish current output data.
Dollars per 0.1 ounce-per-minute of speed is only shown here where the output number belongs to the machine actually being sold.
Pump
Price
Verified output
$ per 0.1 oz/min
Spectra S1 Plus
$226
0.55 oz/min (current model, verified)
$41
Momcozy S9 Pro
$119.99
Not published for this model
Not calculable
Elvie Stride 2
$269.99
Not independently measured for this model
Not calculable
Medela Pump In Style, Handsfree
$130.99
Not published for this model
Not calculable
Three empty cells is the finding, not a formatting error. For reference only, not a fair comparison. It prices machines nobody is currently selling. The discontinued-configuration Medela worked out to roughly $108 per 0.1 oz/min. The predecessor Momcozy S9 worked out to roughly $60. Neither number describes anything you can buy today, so neither belongs in the table above.
Is either company’s “recall” something to worry about?
Momcozy’s wearable pumps, including the S9, S9 Pro, S12, S12 Pro and M1, appear on a Health Canada recall notice. That notice is dated January 28, 2023. Read the notice itself before reacting to the word: the hazard listed isn’t a burn, fire or suction failure. It’s that the devices were sold in Canada without the required medical device license. Health Canada’s own update on the same notice states a license was issued on May 3, 2023, resolving it. That’s a licensing gap closed more than three years ago, not an open safety recall, and I found no U.S. CPSC recall for Momcozy, Spectra, Elvie or Medela breast pumps as of publication.
Does insurance change any of this?
It’s most of the reason the Medela finding above exists at all. Most U.S. health plans are required to cover a breast pump. That coverage is typically supplied through a durable medical equipment provider rather than sold over the counter. The Department of Labor’s guidance on pump-related workplace rules, revised as recently as January 2026, still points to that requirement. That insurance channel is exactly where the older corded Medela has retreated to. Meanwhile general retail moved on to the Handsfree version. If your plan covers a pump, check your supplier’s exact model name and number against what’s reviewed here. Do this before assuming they match.
So which breast pump should you buy?
The Spectra S1 Plus, at $226, has 250 mmHg of suction and 0.55 ounces a minute. Those are numbers BabyGearLab actually verified. It’s not the cheapest pump here, and it isn’t trying to be. It’s the only pump here where the review’s suction number and the box’s suction number are still the same claim. That claim was checked by someone other than the company that sells it.
The Momcozy S9 Pro is the honest budget pick, not because its performance is proven. It earns that because its price and battery life are real: $119.99 and roughly 4.5 hours between charges. That comes with the explicit understanding that you’re trusting a spec sheet with no lab behind it.
The Elvie Stride 2 is the one to want and the hardest to fully recommend at $269.99. The claimed suction jump is real on paper and unconfirmed everywhere else. And the Medela sitting at the bottom of the tally isn’t a bad machine. It’s a case study in a brand name outliving the specific product it was built on. That makes it the loser here on paperwork rather than on air.
What Should You Do Before Buying a Breast Pump?
Two things to do before you buy any of these. Get the exact model number, not just the brand name, from whoever is selling it to you. Insurance supplier included. Suction may be the number driving your decision. If so, buy the pump somebody other than its own manufacturer has actually measured.
Some links on our site may earn us a commission at no extra cost to you. Learn more.
Updated August 2026: scored on arithmetic run against each maker’s own published figures. Dimensions multiplied out and warranty text read to the exclusions. Stated weights were set against two safety bodies that disagree about how heavy a loaded pack may be. L.L.Bean tops the tally; JanSport is second.
In JanSport vs L.L.Bean, the L.L.Bean Original Book Pack wins this tally at 79.6 out of 100. That’s because JanSport’s warranty page says the quiet part in its own words: ‘Unfortunately, this does not mean your lifetime’. The lifetime in JanSport’s lifetime warranty is the bag’s. The same page excludes ‘normal wear and tear’, which is the only way a school backpack has ever died. That is the single fact most likely to change what you buy. It moves the L.L.Bean Original Book Pack to the top of this tally at 79.6 out of 100. The Pottery Barn Kids Mackenzie finishes last at 15.0.
I carried a school backpack myself as a kid, and the zippers gave out before the fabric did. So when a warranty page quietly excludes normal wear and tear, that’s the part that fails first. A lifetime warranty on the bag doesn’t help if that’s the part it skips.
JanSport vs L.L.Bean: Whose Backpack Capacity Math Checks Out?
JanSport’s SuperBreak Plus advertises 1,587 cubic inches (26 L) of capacity. Its own 16.5-by-12.5-by-5.5-inch dimensions multiply out to only 1,134 cubic inches (18.59 L). That’s a claim 40 percent larger than the box its own numbers describe.
Start with the number every backpack comparison ranks on, because it does not survive contact with a calculator.
Each of these four publishes exterior dimensions on its own product page. Multiply height by width by depth and you get the volume of the box the bag occupies. Two of the four also publish a capacity figure. Here is what happens when you put those side by side, at 61.02 cubic inches to the litre.
JanSport vs L.L.Bean: How Do the Published Dimensions Compare?
JanSport prints a capacity 40 percent larger than the box its own dimensions describe. L.L.Bean prints one 2 percent smaller. Same unit, two different things being measured, and neither page says which.
The likeliest explanation for JanSport’s gap is that a soft pack’s stated depth is measured flat and empty. The thing bulges when you load it. That is a reasonable way to measure. It is also not the way L.L.Bean measured. The two numbers sit in the same column on every comparison chart on the internet as if they were.
Pottery Barn Kids is the only one here that publishes an interior measurement as well as an exterior. Its inside is 88.8 percent of its outside, 1,444 cubic inches against 1,626. That is the only calibration between box and usable space any of these four gives you. It comes from the brand people buy for the patterns.
Why doesn’t cheaper capacity per dollar win this backpack comparison?
The Lands’ End ClassMate Large leads at 0.60 litres per dollar. The L.L.Bean Original Book Pack follows at 0.55. But more room for less money is a liability rather than a virtue. That’s true once a small child has to carry what fits in it.
Now the money, on the one basis available for every pack: the exterior box, computed from each maker’s own dimensions rather than from a capacity figure that means different things at different brands.
Pack
Maker’s price
Computed volume
Litres per dollar
Dollars per litre
Lands’ End Kids ClassMate Large
$54.95 (see note)
33.18 L
0.60
$1.66
L.L.Bean Original Book Pack
$44.95
24.71 L
0.55
$1.82
JanSport SuperBreak Plus
$55.00
18.59 L
0.34
$2.96
Pottery Barn Kids Mackenzie Large
not retrievable
26.64 L
—
—
Two notes on the cells that are not clean. Lands’ End’s own product page rendered three price figures at once: $54.95, $46.99 and $18.79. Nothing marks which applies to which colour. I used the highest as the list price. At $46.99 the ClassMate goes to 0.71 litres per dollar and $1.42 a litre. That widens its lead rather than closing it. Pottery Barn Kids renders a range of $39.50 to $89.50 across the whole Mackenzie size run. It publishes no numeral I could tie to the Large. So its row is empty with the reason printed rather than filled with a midpoint.
L.L.Bean gives you 1.63 times the volume per dollar that JanSport does. That is a real spread and it is the kind of number this site normally puts a winner’s rosette on.
Here it points the wrong way, and the next section is why.
More room in the bag is only an advantage until the child has to carry what fits in it.
How heavy should a loaded school backpack be?
The National Safety Council recommends a loaded backpack weigh no more than 5% to 10% of a child’s weight. Children’s Hospital Los Angeles cites the American Academy of Pediatrics’ looser 10-to-15-percent guideline. On a 60-pound child, that’s a swing from 3.0 to 9.0 pounds of allowed weight.
Two credible bodies set a ceiling of no more than 5% to 10%. That’s for how heavy a loaded school backpack should be, and they do not agree on which number is right.
The National Safety Council recommends ‘a backpack weigh no more than 5% to 10% of a child’s weight’. Children’s Hospital Los Angeles, citing the American Academy of Pediatrics, puts it at 10 to 15 percent. It also warns that packs which are too heavy ‘can cause pain, lead to serious injury, and affect posture’. I am not going to pick between them. The point is the size of the disagreement.
Take a 60-pound second-grader. Five percent is 3.0 lb. Fifteen percent is 9.0 lb. The ceiling itself varies by a factor of three depending on which organisation you read. The roundups I went through while checking this quote one of the two figures. They do not mention that the other exists.
How Much of the Weight Limit Does the Empty Bag Use?
Now put the empty bag inside that budget. JanSport publishes 14 oz for the SuperBreak Plus, which is 0.875 lb. L.L.Bean publishes 1 lb for the Original Book Pack. Against the strictest reading, the bag alone spends 29.2 percent and 33.3 percent of the allowance. That’s before a single book goes in. Against the loosest, 9.7 and 11.1 percent. Neither Pottery Barn Kids nor Lands’ End publishes a weight for the packs I read. So I could not run the same arithmetic on them.
And that is why the litres-per-dollar ranking above is the wrong way round for this one product. The Lands’ End ClassMate Large wins it with 33.18 litres of room. A 60-pound child has somewhere between 3 and 9 pounds of allowance to fill it with. Capacity per dollar is a virtue on a cooler and a liability on a first-grader’s back. This is the rare category where the cheapest volume is the thing to be careful about.
Does JanSport’s Lifetime Warranty Beat LLBean’s?
JanSport's lifetime warranty covers the bag's lifetime, not yours. It excludes 'normal wear and tear' — which is how a school backpack actually dies. The page warrants products ‘to the original owner against manufacturing defects in materials and workmanship for the lifetime of the product,’ then defines that term itself: ‘Unfortunately, this does not mean your lifetime. Inevitably, well used (and well loved!) products will eventually wear out.’
Three conditions sit underneath it. The warranty ‘does not cover damage caused by accident, improper care, negligence, normal wear and tear, or the natural breakdown of colors and materials over extended time and use’. It covers the original owner only. And to claim it: ‘All shipping to Warranty must be pre-paid and insured,’ with the further requirement that ‘all items accepted for repair must be clean’. Turnaround is stated as three to four weeks, with free return shipping.
Read those together and the promise narrows a long way. A blown zipper in third grade is covered if it is a defect and not covered if it is wear. The pack has to be washed and shipped at your cost to find out which. That is a real warranty and it is better than nothing. It is not the thing the word lifetime does in a parent’s head in August.
JanSport vs L.L.Bean: Whose Return Terms Hold Up?
L.L.Bean's guarantee used to be the argument-ender in this category. It changed: since February 2018, returns are covered only within one year of purchase. The current text: ‘If you are not 100% satisfied with one of our products purchased directly from L.L.Bean, you may return it within one year of purchase for a refund’. After that window, ‘we will consider any items for return that are defective due to materials or craftsmanship’. Proof of purchase is required. The page also notes that ‘products purchased before February 9, 2018, are not subject to this one-year time limit.’
That date is the whole story. A pack bought before 9 February 2018 carries the old unconditional promise. One bought now carries a year of no-questions returns and a defect review afterwards. Excessive wear and tear is excluded, same as JanSport.
So the two warranties converge more than their marketing suggests. Both are readable, both survive past year one on defects, and both exclude wear. Both also put the burden of proof and, in JanSport’s case, the postage on you.
Pottery Barn Kids and Lands’ End publish no warranty or guarantee text on the product pages I read. That’s true for both the Mackenzie Large and the ClassMate Large. Both companies have returns policies; neither puts a term on that page where a parent comparing packs would see it.
JanSport vs L.L.Bean: How Is This Comparison Scored?
L.L.Bean’s Original Book Pack tops this tally at 79.6 out of 100. That’s driven mostly by a warranty axis worth 30% of the score and a capacity-per-dollar axis worth 25%. Pottery Barn Kids’ Mackenzie Large finishes last at 15.0, largely for publishing fewer of the numbers needed to score it.
Warranty as written — 30%. 25 points each for a duration stated on a page you can read before buying, coverage that survives past one year, no requirement to pay shipping to claim, and wear and tear included. The heaviest axis because it is the only thing that separates a $45 pack from a $55 one over six school years.
Volume per dollar — 25%. From the computed table above, scored against the best. Deliberately not the heaviest axis, for the reason set out two sections up.
Empty weight against the load ceiling — 25%. The lightest published empty weight divided by this one. A pack that publishes no weight scores zero, because the arithmetic cannot be run.
Capacity disclosure — 20%. 100 for a published capacity that matches the maker’s own dimensions, 75 for publishing interior and exterior measurements without a capacity, 40 for a capacity that contradicts the maker’s own dimensions, 25 for dimensions only.
How Do the Four Backpacks Score on Each Axis?
Pack
Warranty 30%
Volume/$ 25%
Empty weight 25%
Disclosure 20%
Total
L.L.Bean Original Book Pack
50
91.0
87.5
100
79.6
JanSport SuperBreak Plus
50
56.0
100.0
40
62.0
Lands’ End Kids ClassMate Large
0
100.0
0
25
30.0
Pottery Barn Kids Mackenzie Large
0
0
0
75
15.0
L.L.Bean’s 79.6 is (50 × 0.30) + (91.0 × 0.25) + (87.5 × 0.25) + (100 × 0.20) = 15.00 + 22.75 + 21.88 + 20.00. Both warranty scores land on 50 for different reasons. JanSport fails the shipping test and the wear test. L.L.Bean fails the wear test and could not be scored on shipping because its page does not say.
Be clear about what the bottom two rows are measuring. Pottery Barn Kids and Lands’ End are not scored down for making bad backpacks. They are scored down for publishing three fewer numbers than their competitors, on three axes that need those numbers. You have the price and the weight in front of you in a store. Use them to re-run the bottom half of this table yourself, and you’ll get a different answer.
L.L.Bean Original Book Pack: 24 L Capacity, Verified
The L.L.Bean Original Book Pack tops this comparison at 79.6 out of 100, and it wins largely on being checkable: a stated 1,476 cu. in. (24 L) capacity you can verify against the bag's own dimensions. That sentence is on L.L.Bean’s own page. It is 97.9 percent of what the pack’s published dimensions multiply out to. That’s the only capacity figure in this field that agrees with the rest of the same page. Add $44.95, a published 1 lb empty weight, and 24 litres of room. There’s also a guarantee whose terms are written down where you can find them. That takes three of four axes.
What it does not have is a laptop sleeve. L.L.Bean’s description does not mention one, and for a school that has gone Chromebook that is a real omission at this price.
Skip this if a padded device sleeve is non-negotiable, or if you bought your last one before February 2018 and expect the same terms.
JanSport SuperBreak Plus: 14 oz Empty Weight
The JanSport SuperBreak Plus is second at 62.0. It wins the axis that matters most for a small child. At 14 oz it is the lightest published empty weight here. That leaves more of a 3-to-9 pound allowance for books. It also carries a 16 by 12.5 inch padded laptop sleeve, which L.L.Bean does not.
It loses on the two numbers this page is built from. That $55.00 buys just 18.59 litres of box, the least volume per dollar in the field by a wide margin. The 26 L on the tag is 1.40 times what its own dimensions describe. The colourway I opened on jansport.com was out of stock with a notify-me form, which is a colour-level fact rather than a line-level one.
Skip this if you are buying on the word lifetime. Read the exclusions first and then decide whether $10 more than the L.L.Bean is buying you anything.
Lands’ End ClassMate Large: Sized for Kids 4 ft 3 in and Over
The Lands' End ClassMate Large is rated best for kids 4 ft 3 in and over. It’s third overall at 30.0 — the most interesting pack here for a reason the tally can't see. Lands’ End is the only maker in this field selling one design in small, medium, large and extra large. Its page sizes them by the child rather than by the grade: the Large is ‘sized for taller/older elementary school kids, it’s best for kids 4’3″ and over’. That solves the actual problem with one-size packs on small torsos.
It also has the most room per dollar on this page, 33.18 litres for a list $54.95. The section above explains why I would not treat that as a recommendation for a young child. Where it falls down is disclosure: no capacity, no empty weight, no warranty on the product page. It also has three price figures rendering at once with nothing to say which is live.
Skip this if your child is under 4 ft 3 in — buy the smaller ClassMate instead of this one, which is Lands’ End’s own advice and the whole point of the sizing.
Pottery Barn Kids Mackenzie Large: Interior Dimensions
The Pottery Barn Kids Mackenzie Large is the named loser at 15.0. The score is almost entirely about what is missing from the page, not what is wrong with the bag. Its interior is listed at 12.5″ wide x 7″ deep x 16.5″ high. Pottery Barn Kids publishes no capacity, no weight, no warranty term, and no price for the Large size. The collection renders $39.50 to $89.50 across every size from Pre-K to rolling.
It earns the single highest disclosure sub-score of anything here that is not a capacity figure. That claim is earned honestly. It is the only pack in this comparison that tells you its interior dimensions as well as its exterior. 12.5 by 7 by 16.5 inches inside, 12.75 by 7.5 by 17 outside. That is 88.8 percent. It is the only evidence on this page of how much of a backpack’s outside is actually usable. Every other brand leaves you to guess.
Pottery Barn Kids also states an age rather than a height: ‘recommended for children ages 6-7 years’ and for those ‘over 48″ tall’.
Skip this if you are optimising cost per year. There is no published warranty term to optimise against, and the patterns are what the premium buys.
CPSC Recall Record: No Open Recalls for These Backpacks
I searched the CPSC recall record for all four brands. I found no open recalls for any of them. Nothing covers the SuperBreak Plus, the Original Book Pack, the Mackenzie or the ClassMate. The hits that do come back for these companies are old and unrelated. L.L.Bean recalled backpack child carriers in 1999. Pottery Barn Kids recalled outdoor trunks in 2005 and wooden pull toys in 2001. Lands’ End recalled children’s zipper pulls in 2005 and pyjamas in 2015. A child carrier is a frame you put a baby in, not a school bag.
Prices are approximate, current as of publication, and change frequently — confirm the latest price with the retailer before buying.
JanSport vs L.L.Bean: Which Backpack Is the Best Buy?
Carrying a school backpack as a kid taught me zippers fail first, so the L.L.Bean Original Book Pack gets my vote, because its warranty reads clearly. The L.L.Bean Original Book Pack is the best long-term buy at 79.6 out of 100. Every figure above came from a maker’s own page or a named safety organisation. Where a maker publishes nothing, the cell says so instead of guessing.
Two things to carry into the store. Read the word lifetime to the end of its own paragraph. JanSport’s ends with normal wear excluded and the postage on you. And find out whether the ceiling you are packing to is 5 percent of your child’s weight or 15. Two credible organisations publish both, and the difference on a 60-pound kid is six pounds of books.
Some links on our site may earn us a commission at no extra cost to you. Learn more.
Updated August 2026: Cuisinart marks the ICE-70 discontinued, selling only spare parts for it; SharkNinja’s current lineup has replaced the CREAMi NC301 with the NC351PU and the Deluxe NC501 with the NC601, both at new prices. Every price re-verified against the maker’s own storefront, cost-per-quart recalculated to include the parts each warranty excludes as wear items, and a genuine conflict between two independent labs over the Whynter’s churn time is reported rather than resolved.
The Cuisinart ICE-21 makes a quart of ice cream for 97 cents. The Whynter ICM-201SB makes the same quart for $4.31. Both figures are the machine itself, spread across three years of making two quarts a month, and the ingredients are identical either way.
That is the whole decision, and the $240 in the middle of it buys exactly one thing: not having to plan a day ahead. The Ninja CREAMi is the loser here on every axis except speed, and even that comes after a 24-hour freeze.
What does a quart actually cost on each of these machines?
Seventy-two quarts over three years, which is two a month. The right-hand column adds the parts each maker’s warranty specifically refuses to cover, because those are the parts you buy twice.
Machine
Price, from the maker’s own store
Machine only, per quart
Parts the warranty excludes
Per quart with those parts
Cuisinart ICE-21
$69.95
$0.97
Replacement freezer bowl, $37.95 on Cuisinart’s own parts listing
$1.50
Cuisinart ICE-70P1 (2 qt)
$189.95
$2.64
not separately priced for this model
—
Ninja CREAMi NC351PU (7-in-1)
$199.99
$2.78
Pint containers, 4-pack, $29.99
$3.19
Ninja CREAMi NC601 (Deluxe, 11-in-1)
$249.99
$3.47
Deluxe tub, single, $16.99
$3.71
Ninja CREAMi Swirl NC701
$299.99
$4.17
Pint containers, 4-pack, $29.99
$4.58
Whynter ICM-201SB
$309.99 on sale, $339.99 list
$4.31
none published
$4.31
The arithmetic is all division by 72. $69.95 gives $0.9715. Add the $37.95 bowl and $107.90 gives $1.4986. $199.99 gives $2.7776, and $229.98 with the pints gives $3.1942. $309.99 gives $4.3054. Run it on your own frequency and every one of these numbers moves.
Two honest gaps in that table. The $37.95 freezer bowl is the price Cuisinart lists on its ICE-70 spare-parts page; I could not confirm a separate published price for the ICE-21’s bowl, so that row assumes the two cost the same, which they may not. And Whynter’s own storefront wants $309.99 while Home Depot has been listing the same model at $259.99, which would drop it to $3.61 a quart. The spread between retailers on this one machine is wider than the entire price of a Cuisinart.
Prices are approximate, current as of publication, and change frequently — confirm the latest price with the retailer before buying.
Which parts will you be buying again, and which ones does the warranty refuse?
Warranty length is the number every listing prints. What the warranty covers is in a separate document, and on one of these machines the two barely overlap.
What the listing says
What the actual terms say
Verdict
Ninja: 1-Year Limited Warranty
SharkNinja’s warranty document excludes normal wear on “containers, lids, paddles, and motor bases,” is non-transferable, and charges the owner a $19.95 to $20.95 return-shipping fee on a covered repair
It is a year on the motor, not on the parts owners report breaking, and it is not free.
Ninja, implied: keep it a year and you are covered
The same document lists as a void condition “failure to keep the well of the motor base clear of food spills and other debris”
A housekeeping clause. Read it before you call support about a CREAMi that has had ice cream spilled in it.
Cuisinart: 3-year limited warranty
Covers defects in materials and workmanship under normal home use. Excludes normal wear and tear, commercial use, and labor and installation costs. Silent on the freezer bowl and silent on transferability.
Three years is real and it is triple Ninja’s. What it does for the one component that wears out is unstated in Cuisinart’s own posted terms.
Whynter: 1 year, parts and labor
Original owner only, non-transferable, limited to the 48 contiguous states, freight and insurance excluded
Alaska, Hawaii and anyone buying secondhand are outside it entirely.
Ninja: “1 pint” capacity
Ninja’s owner guides say only “DO NOT fill containers past the MAX FILL lines” and publish no ounce figure for the 16 oz pint or the 24 oz Deluxe tub
The usable volume of a CREAMi container is a moulded line, not a number the maker will put in writing.
Roundups still recommending the NC301 and NC501
Neither appears as a standalone product on SharkNinja’s current ice cream maker page. The 7-in-1 slot is NC351PU at $199.99, the Deluxe slot is NC601 at $249.99. Both old pages resolve, out of stock.
A live price on a dead model number. Check the maker’s category page, not the product page.
Anyone quoting $189.95 for a “Cuisinart ICE-70”
Cuisinart marks the ICE-70 Discontinued and out of stock, selling only a lid, a paddle and a freezer bowl for it. The live 2-quart machine is the ICE-70P1.
The price is real. It belongs to a different SKU.
Owner reports line up with that first row uncomfortably well. On the CREAMi Deluxe listing at Home Depot, 6 of the 30 reviews rendered on the page I read described a hardware failure: a pint bowl cracked on first use, stripped mixer threads, a paddle that came off the drive shaft after two years, a unit overheating under the blade inside a year, and an outer lid broken within two weeks. On the equivalent Lowe’s listing, 6 of the 20 reviews on that page said much the same, including one owner on their fifth machine in three years. On a Sam’s Club page, none of the 10 reviews mentioned a failure.
Those are page samples, not the full review counts, and I am labelling them that way on purpose. The Home Depot listing carries 320 reviews in total and the Lowe’s listing 318; I read the ones the page served, not all 638. For the Cuisinart ICE-21 and the Whynter I could not get a comparable sample at all — both listings return an aggregate rating, 4.8 stars across 3,683 reviews and 4.5 across 86 respectively, but the review text itself would not load. So there is no owner-reported failure count for those two machines in this piece, and inventing one would be worse than saying so.
Two labs timed the same Whynter and came back 34 minutes apart
America’s Test Kitchen, in a review updated 11 June 2026, put the Whynter ICM-201SB at an average of about 26 minutes and called the output supersmooth, rating the machine Excellent. Reviewed, updated April 2025, reported the same model taking over an hour, its paddles failing during mix-in additions, and a berry frozen yogurt still too thin after 90 minutes.
Same machine. Two named, dated, professional outfits. A factor of three between them. Neither one is dismissible and I am not going to pick the flattering number for you.
The other two machines drew no such disagreement. ATK measured the ICE-21 at about 17 minutes, rated it Excellent and described the results as even-textured and velvety. It measured the CREAMi’s process cycle at about 1 minute 40 seconds and rated the machine Mixed: raspberry sorbet came out well, vanilla ice cream came out soft and liquidy on a conventional recipe. That last point matters more than it sounds. The CREAMi wants recipes written for the CREAMi.
How much electricity does any of this use? Under a dollar in three years.
The CREAMi draws 800 watts but runs for about two minutes. Seventy-two quarts means 144 pints, or roughly 4.8 kilowatt-hours, or 88 cents. Both machines cost less to run for three years than one pint of the ice cream they make. Ignore the wattage. The cost here is the box and the plastic.
The tally: wait, price per quart, texture, warranty, bulk
Three machines, five axes, weights I chose and will defend in a sentence each. Scores out of 10.
Axis (weight, and why)
Cuisinart ICE-21
Ninja CREAMi NC351PU
Whynter ICM-201SB
Time from deciding to eating, 25%. This is what decides whether a machine gets used or shelved.
4 — the bowl has to be frozen solid first, and Cuisinart publishes no pre-freeze time at all on the product page. Then about 17 minutes.
3 — 24 hours to freeze a pint, per Ninja’s owner guide, then 1 to 2.5 minutes
9 — no pre-freeze of any kind; 26 minutes by ATK, over an hour by Reviewed
Cost per quart with warranty-excluded parts, 25%. The figure you pay, not the figure on the shelf.
10 — $1.50
4 — $3.19
3 — $4.31, or $3.61 at Home Depot’s price
Texture in independent testing, 20%. The reason to make it at home instead of buying it.
7 — ATK Excellent and supersmooth, marked down for Reviewed’s thin results and paddle failures
Warranty on the part that breaks, 15%. Length is easy; coverage is the question.
5 — three years, but silent on the bowl
2 — one year, non-transferable, containers and paddles excluded, $19.95 return shipping, housekeeping void clause
3 — one year, non-transferable, 48 states, freight excluded
Bulk and storage, 15%. Every one of these has to live somewhere.
8 — 7.4 lbs, but the bowl claims freezer space permanently
6 — 15.9 lbs, plus two or more pints frozen at all times
3 — 24 to 28 lbs depending on which spec sheet you read; it wants a permanent counter
Weighted tally
7.3
3.8
5.3
The Cuisinart’s 7.3 is 1.00 from wait, 2.50 from price, 1.80 from texture, 0.75 from warranty and 1.20 from bulk. Nothing in that spread is close.
Now move one weight. Push the wait to 45 percent and drop the price to 15, which is exactly how someone who already knows they will not plan ahead should score this, and the Whynter takes it, 6.5 to 6.4. The CREAMi goes to 3.6 and stays last on every weighting I tried. That is the useful result: the argument here is between the cheapest machine and the fastest one, and the CREAMi is not in it.
Every timing above belongs to the lab that ran it, named in the sentence. No machine here passed through our kitchen. Our editorial standards set out how that works.
Cuisinart ICE-21: the cheapest quart, on the longest wait
Seventy dollars, 1.5 quarts a batch, three-year warranty, 7.4 pounds. ATK has named it the canister-style machine to beat for years, and the design is why: there is almost nothing in it to fail. A motor turns a paddle inside a bowl full of frozen coolant.
About 17 minutes of churning, once the bowl has been frozen. The freezing is the part nobody schedules.
The bowl is also the catch, twice over. It has to be frozen solid, and Cuisinart’s own product page will tell you the churn takes 20 minutes or less while saying nothing whatsoever about how long the freezing takes. That is a strange thing to leave off a spec sheet for a machine whose entire workflow depends on it. The practical answer is that the bowl lives in your freezer permanently, which costs you the space of a large mixing bowl for as long as you own it.
Second catch: one bowl, one batch a day. The bowl thaws while it churns.
Skip the ICE-21 if your freezer is already full, or if the reason you want an ice cream maker is that somebody asked for ice cream this evening.
Ninja CREAMi: is it an ice cream maker or a frozen-block processor?
The second one, and that is not a criticism so much as a category correction. The CREAMi does not churn a liquid while freezing it. It shaves and re-blends a block you froze yesterday, which is why it can turn a protein shake, cottage cheese or skim milk into something scoopable — bases a churner would render into ice chips.
That is a real capability and nothing else in this comparison has it. It is also the entire case for the machine, because on price, texture with conventional recipes, warranty coverage and owner-reported durability it finishes last.
A pint holds two generous servings, which is why the four-pack of spare containers is not really optional.
Buy the current model numbers if you buy at all. NC351PU is the $199.99 7-in-1 and NC601 is the $249.99 Deluxe; the NC301 and NC501 that most roundups still name are out of stock on SharkNinja’s own store and gone from its category page. The Swirl NC701 at $299.99 adds a soft-serve dispenser and is the only one of the older numbers still current.
Skip the CREAMi if you want conventional custard-base ice cream, or if the household is not already in the habit of prepping food a day ahead. It is a meal-prep appliance that makes dessert.
Whynter ICM-201SB: what does no planning cost?
About $240 over a Cuisinart, and 20-odd pounds of permanent counter space. The Whynter carries its own compressor, so there is no bowl to freeze and no pint to plan. Pour in a chilled base, come back, then start another batch immediately if you want one.
Whether it does that in 26 minutes or 90 is the open question above, and it is a real one, because the fast figure is most of the argument for the machine. The other thing worth knowing before you spend: the warranty is one year, does not transfer, and does not apply outside the 48 contiguous states.
Skip the Whynter if ice cream is a two-or-three-times-a-summer event. At that rate the compressor is paying for a convenience you use six times.
Is any of this recalled?
No. A search of the Consumer Product Safety Commission’s recall database returns no recall on any ice cream maker from SharkNinja, Cuisinart or Conair, or Whynter.
Two adjacent recalls exist on these same corporate names and neither touches these machines. Consumer Reports covered SharkNinja’s May 2025 recall of about 1,846,400 Ninja Foodi pressure cookers over a lid that could open under pressure, with 106 reported burns. And CPSC expanded a recall of Cuisinart grill brushes on 27 August 2026 to more than 3.6 million units over metal bristles coming loose. Different products, different factories. They are here because you would want to know I looked.
Where the tally and my own money disagree
The tally says Cuisinart ICE-21 at 7.3. Re-weighted for someone who values not planning, it says Whynter at 6.5 against 6.4, which is a real tie and the only honest place to break one here.
I would still buy the ICE-21, and the reason is the disagreement in the section above. The Whynter’s case rests on speed. One of the two independent labs that timed it says it is fast and one says it took over an hour and the paddles failed. Spending $240 extra on a convenience that two professional kitchens cannot agree exists is a bet, not a purchase.
Buy the $70 Cuisinart. Keep the bowl in the freezer so the wait is already spent. If you make enough ice cream to outgrow it, you will have learned exactly which constraint annoys you, and you will be spending the upgrade money on evidence instead of a guess.
One more thing to check before you order
Open the maker’s own category page, not the product page, and confirm the model number you are about to buy is still on it. Two of the six machines here failed that check, and both still have live pages with live prices.
Some links on our site may earn us a commission at no extra cost to you. Learn more.
Updated August 2026: every price re-verified against each maker’s own current storefront. Fuel cost is computed from Oak Ridge National Laboratory’s instrumented 60 mph results at the EIA’s current $4.085-a-gallon average. The SkyBox NX’s latch-and-lock failure pattern, reported by four independent sources, is covered in full. Yakima’s own instruction sheet lists a torque wrench among the SkyBox NX’s mounting contents; SportRack’s mounting hardware is tool-free.
Thule vs Yakima vs SportRack comes down to 75 points out of 100. The Thule Motion 3 XL wins, and the reason is not capacity. It is the latch. Yakima’s SkyBox NX finishes on 57.5 because four independent outlets and its own REI reviewers report the same failure. The problems repeat: handles that stick, latches that need several attempts, and key cores that ice up. In one owner’s account, the box opened on the freeway. The SportRack Vista XL takes 49.5 and costs $399.99, less than half the Thule’s $919.96. Before any of that matters, check one number: your crossbar spread.
On a long road trip, gear rattling above your head for hours is something you notice fast. I have driven enough of them to know the rule. A latch that needs a few tries to close is a problem before you leave the driveway. So I’d rather pay more for hardware I trust at highway speed than save a little and worry.
Thule vs. Yakima vs. SportRack: which cargo box wins?
After long road trips with gear lashed to the roof, the Thule Motion 3 XL is my pick, because its latch is the failure I won’t risk. The Thule Motion 3 XL wins this comparison at 75 out of 100 points. Yakima’s SkyBox NX finishes at 57.5 points on a reported latch-failure pattern. The SportRack Vista XL scores 49.5 points but costs $399.99, less than half the Thule’s $919.96.
Best overall: Thule Motion 3 XL, $919.96 from Thule. 18 cu ft, dual-side, and the widest fit range of the three.
Best budget, if it fits: SportRack Vista XL, $399.99 at Home Depot — $50 below the maker’s own price. Rear-opening only, and a six-inch fit window.
The one I would not buy: Yakima SkyBox NX. Good shape, good access, and a latch record I cannot argue past.
How much does a rooftop cargo box cost you in gas?
Oak Ridge National Laboratory tested two vehicles with a rooftop box mounted at 60 mph. A 2009 Ford Explorer went from 24.9 mpg to 22.9 mpg. A 2009 Toyota Corolla went from 42.5 mpg to 33 mpg. That works out to $14.33 extra per 1,000 highway miles for the Explorer and $27.67 for the Corolla.
Which one fits your roof comes down to your crossbar spread. The published clamp range differs for the Thule Motion 3 XL, the Yakima SkyBox NX, and the SportRack Vista XL. The range comes from real work, so use the work instead.
This comes from SAE paper 2014-01-1614, by John Thomas, Shean Huff and Brian West at Oak Ridge National Laboratory. They ran two vehicles at a steady 60 mph with and without a rooftop cargo box. A 2009 Ford Explorer went from 24.9 mpg to 22.9. A 2009 Toyota Corolla went from 42.5 mpg to 33. Those are raw measured pairs, which is what makes them useful — you can compute in either direction from them.
Arithmetic, so you can redo it with your own car and your own pump price: 1,000 ÷ 33.0 mpg = 30.30 gallons, and 30.30 × $4.085 = $123.79. Subtract the 23.53 gallons the same car needs without the box and the difference is $27.67.
How Much Does a 3,000-Mile Trip Cost With a Box On?
On a 3,000-mile summer round trip that is $43 on the Explorer and $83 on the Corolla. The compact car is hit roughly twice as hard, which is the opposite of most people’s intuition about roof boxes and big vehicles.
A vehicle carrying a rooftop box on an open road.
Now the part worth knowing if you ever quote this study, because I nearly repeated the error myself. Oak Ridge’s own summary reports the Explorer as “a drop of 9 percent”. It reports the Corolla as “22 percent.” Those are two different calculations. The Explorer’s 9 is the rise in fuel consumption (40.16 to 43.67 gallons, up 8.7 percent). The Corolla’s 22 is the fall in fuel economy (42.5 to 33 mpg, down 22.4 percent). Compute the Corolla the same way as the Explorer and its consumption rises 28.8 percent, not 22. Every page that prints “9% and 22%” as though they were the same measurement has made an error. It has copied a mixed pair without checking.
Two independent confirmations that the effect is real and large. Consumer Reports tested at 65 mph in 2020. It measured a 2019 Nissan Altima falling from about 48 mpg to about 39 with a rack and carrier. A 2019 RAV4 fell from 41 to 35 in the same test. Alan Meier at Lawrence Berkeley National Laboratory put the national bill at roughly 100 million gallons of gasoline a year. His co-author was Yuche Chen. That is about 0.8 percent of all US light-duty fuel use. Most of it is burned by boxes nobody bothered to take off.
Take it off between trips. That is the cheapest thing you can do for a roof box, and it costs nothing.
What Crossbar Spread Do Thule, Yakima, and SportRack Need?
The Thule Motion 3 XL accepts a 14.75-inch crossbar spread. Yakima’s SkyBox NX XL accepts 16 inches. The SportRack Vista XL takes just 6.1 inches. Each box only clamps within its own published range. Measure across the tops of your bars, centre to centre, before you shop. If your spread falls outside a box’s window, nothing else on this page matters.
Box
Crossbar spread accepted
Width of the window
Source
Thule Motion 3 XL
2113⁄16” to 369⁄16“
14.75″
Not published by Thule. Figure from etrailer’s listing for TH87PN, corroborated at 2113⁄16“–365⁄8” by Better Trail’s own measurements
Yakima SkyBox NX XL
24″ to 40″
16″
Yakima’s own instruction sheet
Yakima SkyBox NX LG
24″ to 36″
12″
Yakima’s own instruction sheet
SportRack Vista XL
235⁄8” to 293⁄4“
6.1″
Public Lands spec table. SportRack does not publish it
Which Column Matters Most Here?
Read the first column, not the third. The Thule accepts bars more than two inches closer together than any Yakima here. That means on a short-roofed car with fixed factory points, the Motion 3 mounts where a SkyBox NX will not. The cheapest box on this page has a six-inch window. That is less than half the Yakima’s, and well under half the Thule’s. The budget option is also the one most likely not to fit your car.
Yakima publishes a hatch-clearance figure too: 55 inches minimum for the XXL. That is measured from the centre of the front crossbar to the edge of the open trunk door. Thule publishes no clearance number for the Motion series at all, and SportRack publishes none either. Two of three tell you to measure and do not tell you what to measure against.
What Do Thule, Yakima, and SportRack’s Warranties Leave Out?
Limited lifetime coverage on a plastic shell sounds like the end of the argument. It is the start of one. I read the actual terms.
Maker
The claim
What the published terms exclude
Verdict
Thule
Limited lifetime, registration not required
Normal wear and tear, scratches, accidents; modifications or unauthorised repairs; damage to non-Thule products used with it; anything traced to misuse, overloading, or not following the written instructions
Real, with the standard carve-outs
SportRack (a Thule Group brand)
Limited lifetime
Everything above, plus no coverage at all on units bought outside the US, Canada and Mexico, and coverage terminates if the purchaser transfers the product to another person
Weakest of the three — it does not survive resale
Yakima
Limited lifetime, stated on every SkyBox NX product page
Could not be read. Yakima’s warranty terms page would not render to anything but a full browser
Unverified
The SportRack clause is the one nobody mentions. Buy a Vista XL used and the lifetime warranty is already over. Sell yours and you are handing the next owner an uncovered box. On a $400 item that may not change your decision, but you should make it knowingly.
How Do Thule, Yakima, and SportRack Score, Axis by Axis?
The Thule Motion 3 XL scores 75 out of 100 across five weighted axes. Yakima’s SkyBox NX scores 57.5, and the SportRack Vista XL scores 49.5. The weights are mine, and they are all written down, so move them and re-run it.
Latch and lock reliability, 30%. The heaviest weight by a distance. Every other flaw on this page costs you money or patience. This one costs you your gear at 70 mph.
Fit range, 20%. A box that does not clamp to your bars scores zero on everything else.
Loading access, 20%. Dual-side against rear-only is a tax you pay at every rest stop, on the traffic side of the car.
Aero and noise, 15%. Held down deliberately, because no one has published a per-box fuel measurement. The ORNL figures above are for a generic box, not for these three.
Price, 15%. A $520 spread across the field is real, and it is still the least important thing about a box bolted over your windshield.
How Is Each Scoring Axis Weighted?
Axis (weight)
Thule Motion 3 XL
Yakima SkyBox NX XL
SportRack Vista XL
Latch and lock (30%)
8 — “occasionally finicky to fully engage” per Better Trail; no owner failure pattern; 4.6/5 from 66 REI reviews
3 — four independent sources on sticking handles, iced key cores and failed latches
6 — owners report the lid catching on the latch; no failure pattern, but no independent testing either
Fit range (20%)
9 — 14.75″ window, lowest minimum spread
7 — 16″ window but a 24″ floor
3 — 6.1″ window
Loading access (20%)
9 — dual-side, one-hand
9 — dual-side, 11″ hatch height
3 — rear opening only
Aero and noise (15%)
7 — “minimal wind noise” per Better Trail; 8.3/10 from GearJunkie’s fuel-consumption route testing
5 — “near-constant hum” and 1–3 mpg lower over 10,000+ long-term miles; 6.8/10 from GearJunkie
4 — no independent testing exists
Price (15%)
3 — $919.96
6 — $679.19 at REI, $849 from Yakima
9 — $399.99, below the maker’s own $449.95
Weighted tally
75.0
57.5
49.5
Worked example: the Thule scores (8 × 0.30) + (9 × 0.20) + (9 × 0.20) + (7 × 0.15) + (3 × 0.15) = 2.40 + 1.80 + 1.80 + 1.05 + 0.45 = 7.50, or 75 out of 100.
Drop the latch weight to 10 percent and spread it across price and access, and the Yakima climbs past the SportRack comfortably. I am not going to do that. Read the next section and decide whether you would.
Yakima SkyBox NX: the same complaint from four directions
Four independent sources — including Yakima's own REI reviewers — report the same pattern: handles that stick, latches that need several attempts, key cores that ice up, and in one case a box that opened on the freeway. That $235 spread on one SKU is worth knowing before you buy anywhere.
The box itself is well shaped. GearJunkie liked the larger nose, which takes duffels and small totes in space most boxes waste. Dual-side opening, an 11-inch hatch height, 165 lb load limit, a full metal handle that locks the gear in and the box to the car.
Then the latch. GearJunkie: the handle can stick when opening, and the latches sometimes need several attempts to engage with the box loaded. Switchback Travel, after well over 10,000 miles across four seasons on three vehicles, reports the updated key core “iced up on us on several occasions,” delaying departures, and concludes that resolving the icing would make the box an across-the-board recommendation. Switchback’s roundup repeats the lock-icing finding independently. And REI’s own reviewers supply the ugly end of it: one owner describes the box flying open on the freeway and losing gear despite appearing locked, another a broken latch at five months, another a side hinge broken at eight.
What Do the SkyBox NX Ratings Show?
The ratings show it. SkyBox NX XL sits at 4.4 from 57 REI reviews, which looks fine. The LG sits at 3.7 from 24, and the distribution is the tell: eleven 5-star, five 4-star, zero 3-star, five 2-star and three 1-star. Eight of twenty-four at the bottom and nothing in the middle. That is not a mediocre product. That is a product that works or fails.
Plenty of roundups still call the SkyBox NX mounting tool-free; Yakima’s own instruction sheet lists a torque wrench among the box’s contents, so that description does not hold. Thule’s PowerClick genuinely is tool-free, with an audible click when the clamp hits torque.
Skip this if you park outdoors anywhere that freezes. The icing report comes from someone who lived with the box through a full winter, and a lock that will not open is a trip that does not start.
Thule Motion 3 XL: the fit range is the real feature
$919.96 from Thule, $919.89 at REI, $1,115.45 at etrailer. The seven-cent gap between the first two is Thule enforcing minimum advertised price; the etrailer figure is $195 over both, so buy from Thule or REI.
18 cubic feet, 490 litres by Thule’s figure, 165 lb load limit, dual-side one-hand opening, PowerClick tool-free clamps with a torque indicator, and a SlideLock that separates locking from opening so the key will not come out of an unlatched box. Thule quotes 51 lb; Better Trail weighed one at 55 lb 14 oz over several months of repeated installs. Believe the scale.
REI reviewers put the XL at 4.6 from 66 and the L at 4.9 from 32, with 28 five-star ratings and not a single review below four. Better Trail found the latching “occasionally a little finicky to fully engage,” the glossy paint prone to showing scratches, and the mounting hardware protruding further into the interior than Yakima’s. Those are complaints about a good box.
Availability, precisely: Thule’s individual Motion 3 product pages read “temporarily out of stock” on 30 August 2026 while the category page listed them as available, and REI has the XL and L on pre-order with black sold out. Actively sold, not discontinued, not necessarily shippable this week.
Skip this if the box goes on in June and comes off in September. Nine hundred dollars of hardware you touch twice a year is money that belongs in the trip.
SportRack Vista XL: Cheap, Rear-Opening, Short on Specs
$449.95 from SportRack, $399.99 at both Home Depot and Public Lands. Retailers undercutting the maker by $50 is unusual and it is in your favour. 18 cubic feet, 63 by 38 by 19 inches, UV-resistant ABS, tool-free mounting hardware, a lock for the contents.
Loading luggage into a vehicle before a trip.
Two numbers I will not print, because SportRack does not publish either and the retailers disagree. Box weight: Rack Attack says 42 lb, Switchback Travel says 28 lb. Maximum load: Public Lands’ spec table says 110 lb, Rack Attack says 165 lb. A 55-pound disagreement on load capacity matters when you are checking it against your vehicle’s roof rating, and the maker is silent. Ask before you buy and get the answer in writing.
Owner feedback is better than the Ledger score suggests: 4.5 from 65 reviews at Public Lands, 135 reviews at Home Depot. The recurring complaints are specific and mild. Manufactured mounting holes line up at the front and not at the back on some racks, requiring drilling. The lid catches on the latch. And reaching gear packed at the front through a rear-only opening is exactly as awkward as it sounds.
There is also no independent testing of this box anywhere I could find. Not GearJunkie, not Switchback beyond a one-line placement at the bottom of a roundup, not Consumer Reports. Its entire evidence base is retailer reviews.
Skip this if your crossbar spread falls outside 235⁄8” to 293⁄4“, or if you need a published maximum load to check against your roof rating before you sign off on it.
Have Thule, Yakima, or SportRack Ever Been Recalled?
No open recall covers the Thule Motion 3, the Yakima SkyBox NX or the SportRack Vista XL.
Two nearby items, since I am recommending two of these brands. Thule’s own recall page lists seven historical actions and none is a cargo box, but one is close: the Tepui HyBox and HyBox Wedge, recalled in May 2021 over the risk of the unit detaching from the vehicle roof. Different product line, different mounting system, and not a Motion 3 issue. Yakima’s recalls are bike racks, including the 2018 HoldUp EVO and SingleSpeed action over snap pins that could work loose. No SkyBox has ever been recalled.
The method, stated so you can weigh it: the CPSC’s recall database blocks automated retrieval, so this came from each maker’s own published recall notices, NHTSA campaign records and Canada’s recall database rather than a direct CPSC query. If you are about to spend $900, spend two minutes checking cpsc.gov and nhtsa.gov yourself before you do.
Should You Buy the Thule, Yakima, or SportRack Box?
The Thule Motion 3 XL wins by 17.5 points on our tally, and it's the one to buy if the budget allows. That is not the interesting part.
The interesting part is the bottom two. Yakima 57.5, SportRack 49.5. Eight points, and a $279 price gap between $679.19 and $399.99. That is close enough to argue about, so here is the argument.
I would take the SportRack, on one condition: measure your crossbar spread first and confirm it lands inside that six-inch window. If it does, you are paying Yakima $279 extra for a box whose one documented weakness is the exact component that keeps your gear on the roof. Rear-only loading is a real inconvenience. It is an inconvenience you experience in a car park, not on a freeway.
If your spread does not fit the SportRack, buy the Thule and skip the middle entirely.
Prices are approximate, current as of publication, and change frequently — confirm the latest price with the retailer before buying.
Some links on our site may earn us a commission at no extra cost to you. Learn more.
A sunrise alarm clock is a convenience product, not a medical device. If you are dealing with insomnia, seasonal low mood, or another sleep or mood condition, talk to a healthcare professional about what is right for you rather than to a lamp.
Updated August 2026: the Philips SmartSleep HF3650 has come off Philips’ own US light-therapy listing page, so it finishes last on the tally. This comparison scores light output, the distance it’s measured at, and what each maker puts in writing about it; the subscription-cost math for Hatch+ now lives on our dedicated Hatch+ page. Prices and warranty terms were re-read at each maker’s own site this month.
The Lumie Bodyclock Shine 300 wins this on a weighted tally, 82.1 out of 100, and the Philips SmartSleep HF3650 loses it at 49.3. But the number worth carrying away is not either of those. It is this: the JALL Wake Up Light, at $34.99 on JALL’s own store, puts 545 lux on a meter twelve inches away. The Hatch Restore 3, at $169.99, puts 285 lux on the same meter at eighteen inches. Six inches of nightstand is worth more than $135 of clock.
Four Sunrise Alarm Clocks, Four Product Pages, One Lux Figure
A sunrise clock has exactly one job that a $12 clock radio cannot do: put light on your closed eyelids before the sound starts. So the first thing to check is what each maker is willing to say about that light. I read all four product pages and the support and warranty documents behind them.
Maker
What it publishes about brightness
What an independent meter recorded
Verdict
Hatch (Restore 3)
No lux, lumen or nit figure on hatch.co. Its own support article on brightness says only to move a slider to your preferred setting.
610 lux at 12 in, 285 lux at 18 in
No claim to check
Philips (SmartSleep HF3650)
Nothing on the US product page. Philips’ own manual for the HF3520, a different wake-up light, prints ‘Approx. 1 lux – 300 lux at 45 cm’.
650 lux at 12 in, 325 lux at 18 in
The only conditioned figure in the field — and it is on another model
Lumie (Bodyclock Shine 300)
No lux or lumen figure on either the US or the UK page. Both publish the sunrise duration range instead.
730 lux at 12 in, 360 lux at 18 in
No claim to check
JALL (Wake Up Light)
No lux or lumen figure on jall-us.com. The listing counts colours and nightlights.
545 lux at 12 in, 275 lux at 18 in
No claim to check
Every measured figure in that right-hand column comes from one place: Optimize Your Biology’s light-meter testing of 15 sunrise clocks, whose author describes the method as putting each clock in front of a spectrometer and measuring lux every six inches across brightness settings. That is a single independent tester carrying the entire evidentiary weight of a product category. More on what that limits at the bottom.
Three of four makers publish nothing. That is not an oversight and it is not shyness about a bad number. Brightness without a distance is not a specification, and a maker that prints one has to pick a distance and defend it.
The gap between the clock and your face is the variable that decides how much of its light reaches you.
Six Inches of Nightstand Is Worth Roughly Twice the Light
The same test publishes readings at two distances I could read consistently across the whole field: one foot and eighteen inches. That gives you something the marketing does not. Divide the eighteen-inch reading by the twelve-inch reading and you get how much of the light survives half a foot of extra nightstand.
If these clocks behaved like a bare bulb, the answer would be fixed. Illuminance from a point source falls with the square of distance, so going from 12 to 18 inches should leave (12 ÷ 18)² = 44.4 percent. Here is what the meter actually recorded.
Clock
Lux at 12 in
Lux at 18 in
Light retained
Point source predicts
Gain from moving 18 in to 12 in
Lumie Bodyclock Shine 300
730
360
49.3%
44.4%
2.03×
Philips Hue Twilight (fifth clock, same test)
840
395
47.0%
44.4%
2.13×
Philips SmartSleep HF3650
650
325
50.0%
44.4%
2.00×
Hatch Restore 3
610
285
46.7%
44.4%
2.14×
JALL Wake Up Light
545
275
50.5%
44.4%
1.98×
All five hold on to more light than a point source would — between 46.7 and 50.5 percent against a predicted 44.4. That is expected physics, not a measurement error. A 2025 paper by Cecilia Stari, Marcos Abreu, Martín Monteiro and Arturo C. Martí, Light illuminance does not always decay with the inverse of the square of the distance (arXiv 2501.00622), sets out the cases: a flat source keeps illuminance nearly constant while you are closer to it than the source is wide, and only at distance does it start behaving like a point. A sunrise clock is a broad diffusing dome rather than a filament, and it is read from one to three feet away. That is the region where the simple rule stops applying.
The practical version fits in a sentence. Move any of these clocks from eighteen inches to twelve and you get about twice the light — 1.98× on the JALL, 2.14× on the Hatch. That swamps every difference between the four. The JALL at twelve inches delivers 545 lux against the Hatch’s 285 at eighteen: 1.91 times the light for 4.86 times less money.
What the test does not publish, and what I am not going to invent, is where each clock falls below a useful level. Optimize Your Biology names roughly 250 lux at the pillow as a working floor and reads out to 36 inches on per-model charts; the two distances I could pull cleanly for all five are the two above. If your clock sits across the room on a dresser, none of these numbers describes your morning.
300 Lux at 45 Centimetres, Printed on a Model Philips Sells Instead
One document in this whole category does the job properly. Philips’ own user manual for the HF3520 wake-up light, dated 2022, states the light output as approximately 1 lux to 300 lux at 45 cm. A number and the distance it was taken at. That is the format a buyer can use.
Forty-five centimetres is 17.7 inches, near enough to the eighteen inches the independent meter used that the two are comparable. Philips claims up to 300 lux on the HF3520; the meter recorded 325 lux on the HF3650 at 18 inches. Different models, so this is not a claim checked against itself — but the one maker that publishes a conditioned figure is also the one whose products land where an outside meter says they land. That is worth something.
It also makes the rest of the field look worse rather than better. Philips proved the number can be printed. Three makers chose not to print it.
$34.99 to $169.99, and What Each Price Buys
Lumie Bodyclock Shine 300, $169.00: the brightest reading in the test, the longest sunrise ramp, and the longest warranty. On pre-order at Lumie as of this update.
JALL Wake Up Light, $34.99: within 10 lux of the Hatch at the same distance, for a fifth of the money. The warranty is a claim you cannot read.
Hatch Restore 3, $169.99: the app, the sound library and the routines. The shortest warranty here and the clock’s own light is third of four.
Philips SmartSleep HF3650: the loser. Second-brightest, shortest sunrise ramp, no price on Philips’ own site, and gone from Philips’ US listing page.
Five Axes, and What Each One Is Worth
Here is the tally, with every weight stated so you can move one and see what happens.
Light at 18 inches — 30%. The heaviest weight because it is the only thing a sunrise clock does that a cheap clock radio cannot. Scored as the measured lux divided by the highest in the field, 360.
Longest settable sunrise — 20%. A slow waker needs a long ramp and cannot add one later. Scored as the longest ramp divided by 90 minutes, the longest available.
Price on the maker’s own site — 15%. Light, deliberately: the spread here is real but the light difference is bigger. Scored as $34.99 divided by the price.
Warranty as published — 20%. Months of coverage you can read before you pay, divided by 36, the longest verifiable term here. A warranty you cannot read scores zero.
Listed on the maker’s own catalogue page — 15%. Not the product page, which can outlive the product. 100 if listed and shipping, 60 if listed on pre-order, 0 if the maker has dropped it.
Clock
Light 30%
Ramp 20%
Price 15%
Warranty 20%
Still listed 15%
Total
Lumie Bodyclock Shine 300
100.0
100.0
20.7
100.0
60
82.1
JALL Wake Up Light
76.4
66.7
100.0
0
100
66.3
Hatch Restore 3
79.2
66.7
20.6
33.3
100
61.9
Philips SmartSleep HF3650
90.3
44.4
0
66.7
0
49.3
The Lumie’s 82.1 breaks down as (100.0 × 0.30) + (100.0 × 0.20) + (20.7 × 0.15) + (100.0 × 0.20) + (60 × 0.15) = 30.00 + 20.00 + 3.11 + 20.00 + 9.00. Run the other three the same way and the order does not change.
Two cells are zeros with reasons rather than measurements. The JALL scores zero on warranty because jall-us.com advertises a lifetime warranty behind an activation link and publishes no terms document I could retrieve; that is a mark against the disclosure, not a claim that JALL refuses service. Philips scores zero on price because usa.philips.com renders no price numeral for the HF3650 and I will not substitute a retailer’s. An empty cell with a printed reason beats a filled one resting on a guess.
The gap from first to second is 15.8 points, which is wider than any weight I would defend moving. No tie-breaker is needed here.
360 Lux and Three Years: the Lumie Bodyclock Shine 300
The Lumie takes four of five axes. It measured brightest of the field at both distances. Its sunrise runs 15 to 90 minutes, a range Lumie prints on its own page and the outside test confirms, and the 90-minute end is genuinely rare. Its warranty is three years, worded on the US page as a 3-year limited warranty and on the UK page as a 3-year warranty, with the same repair-or-replace remedy either way.
Two honest marks against it. Lumie’s US page listed the Shine 300 as pre-order with an expected shipping date of 25 September when I read it, which is why it lost 40 points on the availability axis rather than none. And the same independent test rates its flicker as risky at low brightness, which matters if you use it as a bedside reading lamp as well as an alarm.
Price is $169.00 on Lumie’s US store. The UK store lists the same model at £149.00. I am not converting one into the other — the two stores ship to different places and price independently — but it is worth knowing both figures exist before a reseller quotes you a third.
Skip this if you need the clock in the house this month, or you want it doing double duty as a reading light.
$34.99 Buys 275 Lux: the JALL Wake Up Light
Second place, and the reason is arithmetic rather than affection. At 275 lux at eighteen inches the JALL sits 10 lux behind the Hatch Restore 3 and 85 behind the Lumie, on a meter that cannot tell what anything cost. Its sunrise runs 10 to 60 minutes. It is listed and shipping on jall-us.com at $34.99.
The warranty is the problem, and it is a disclosure problem rather than a coverage one. The site carries a banner inviting you to activate a lifetime warranty. I could not retrieve the terms behind it, so I could not score a duration, and a term you cannot read before buying is worth nothing at the point you are deciding. Adjust that cell upward if you are willing to take the banner at face value. Score the JALL 100 on warranty and it goes to 86.3 and takes first place from the Lumie by 4.2 points. That is the single weight in this tally most worth arguing about.
The build is what $34.99 buys. Plastic, an interface that takes a week, and a brand whose US presence is a Shopify store and an Amazon shelf. Nothing here is precious.
Skip this if you want a maker who will still be answering the phone in five years, or if the clock is a gift and the box matters.
One Year From Registration: the Hatch Restore 3
Third at 61.9, and it loses on paperwork more than on light. Hatch’s own warranty page covers the Restore for one year from the date of registration, not the date of purchase — so the clock a buyer forgets to register is a clock with no term running at all. That is the shortest coverage in this field, attached to the joint-highest price at $169.99.
What it does well is not in dispute. The sound library is the best here, the app builds routines that people actually keep, and the free tier is real: Hatch publishes it as 40 sleep sounds, 20 alarm sounds, 20 lights and five unwind options, with Hatch+ at $4.99 a month or $49.99 a year adding roughly 100 of each. If the subscription is the thing you are weighing, we run that math separately in our comparison of the Hatch+ membership against a $50 sound machine, which puts the Restore 3 against the Yogasleep Dohm and the LectroFan EVO on running cost. This page is about the light.
On the light it is third of four, and 285 lux at eighteen inches is 79 percent of what the Lumie delivers at the same distance for a dollar less.
Skip this if the light is the reason you are buying, or if you will not register a product within a year of unboxing it.
Zero Listings on Philips’ Own Catalogue Page: the SmartSleep HF3650
This is the named loser despite testing second-brightest of the four at 325 lux, which beats the Hatch and the JALL comfortably on light alone.
Then I read the catalogue page rather than the product page. Philips’ US light-therapy listing carries eight products — the EnergyLight HF3318 at $299.99, two goLITE BLU models at $199.99, the briteLITE 6, and four Wake-up Lights including the HF3471 at $129.99 and the Wake-up Light Plus HF3485 at $169.99. The SmartSleep HF3650 is not among them. Its own product page still resolves, still renders the specification list, and still carries a 2-year warranty line, but it renders no price and Philips no longer puts it in the menu.
A product URL that still loads is not evidence a maker still sells the thing. Philips has not announced anything about the HF3650 that I could find; it simply stopped listing it, which is the quieter of the two ways a product leaves. If you want a current Philips, the HF3485 at $169.99 is the one the company is actually selling — and there is no independent lux figure for it, so you would be buying it on faith.
Skip this if you want a warranty you can claim on, which is most people. Buying a delisted model means betting the parts and the support outlast the listing.
One Recall on Record, and It Is Not This Clock
Hatch has a recall in the CPSC record and it deserves stating plainly, because searching the brand will surface it. On 3 July 2024 the agency published recall 24-291: about 919,400 US power adapters sold with Hatch Rest 1st-generation sound machines, plus about 44,352 in Canada, because the plastic housing could come off in the outlet and leave the prongs exposed. The notice records 19 reports of the housing detaching and two reports of a minor shock.
The adapter is model CYAP05 050100U and it shipped with the Rest 1st generation, a baby sound machine. The Restore 3 is not in the recall, and neither is the Rest 2nd generation. I found no CPSC recall touching Philips, Lumie or JALL sunrise clocks.
One Tester, One Unnamed Meter, and No Published Protocol
Every lux figure on this page traces to a single source. Optimize Your Biology bought 15 clocks and metered them; that is more than any of the makers did in public, and it is still one person. The instrument is described as a lab-grade spectrometer and is not named. No protocol is published for how the clocks were aimed, what brightness setting each reading used, or how ambient light was excluded. I could not find a second independent light-meter test of this category to check it against.
So the ranking on light is as good as that one test and no better. It does not compare products this operation has handled — we buy nothing and meter nothing, and every number above is somebody else’s reading or somebody’s own published document, named in the sentence where it appears. Our sourcing rules are on the editorial standards page.
What does not depend on that test is the shape of the finding. The distance arithmetic holds whatever the absolute numbers are, because it is a ratio of two readings from the same meter on the same clock. And the warranty terms, the prices and the catalogue listings came straight from the makers.
Prices are approximate, current as of publication, and change frequently — confirm the latest price with the retailer before buying.
Two Decisions, in Order
Decide where the clock sits before you decide which clock it is. Twelve inches from your face rather than eighteen is worth about double the light on every unit measured, which is more than the entire spread between the cheapest and the brightest.
Then buy on the tally. The Lumie Bodyclock Shine 300 if you can wait for stock. The JALL if you want to find out whether light wakes you before spending $169 on the question. The Hatch Restore 3 if the sounds and the app are the point and you will register it. Not the Philips SmartSleep HF3650, which Philips has stopped listing.
Some links on our site may earn us a commission at no extra cost to you. Learn more.
Updated August 2026: GE rates the Opal 2.0 Ultra at 38 pounds of ice a day, not the smaller Opal 2.0’s 24. Every price re-verified against each maker’s own current listings. The Frigidaire’s measured output is 18 percent under its own claim, and the Opal Ultra’s headline number has never been independently tested at all.
Three countertop nugget machines, three output claims, and only two of them have ever been put on a scale by someone who does not sell them. The GoveeLife Smart Nugget Ice Maker Pro takes the tally at 73.5 out of 100 — it is the only one of the three whose headline number survived an independent timing test, and it carries a three-year warranty against one-year coverage from both rivals. Frigidaire’s EFIC256 finishes second at 58.5 and costs $281.75. The GE Profile Opal 2.0 Ultra finishes last at 40, at $449, with $116 a year in filters and a rated output figure no lab has confirmed.
I go through more ice than most people in a year of hot, humid summers. A machine that runs all day without babysitting matters more to me than it might to someone up north. I’ve used enough countertop ice makers at rentals and friends’ places to know a loud compressor gets unplugged fast. What decides which one earns counter space long-term is whether the numbers on the box hold up, not how it looks in a kitchen photo.
Which Nugget Ice Maker Suits Your Kitchen?
Most ice for the money, and the only verified claim: GoveeLife Smart Nugget Ice Maker Pro, $424.99 on Govee’s own store. Reviewed timed a full 3.5-pound bucket in just over 90 minutes.
Cheapest way into real nugget ice: Frigidaire EFIC256, $281.75 at Home Depot. No filter to buy, ever.
The upgrade, if you want the app and the UV chute: GE Profile Opal 2.0 Ultra, $449 direct from GE. You are paying for polish, not for output you can point at.
How Many Nugget Ice Maker Claims Have Been Measured?
Companies don’t exactly lie about ice output. They publish a number produced under conditions they chose and never repeat. So the useful question is not what the box says. It is whether anyone outside the company has checked.
For these three, the answer is one no and two yeses, and the no belongs to the most expensive machine on the counter.
Machine
Maker’s claim
What an independent lab measured
Gap
Verdict
GE Profile Opal 2.0 Ultra
38 lb / 24 h, 1.6 lb per hour
Nothing. Reviewed tested the Opal 2.0, the 24 lb/day model, at roughly 1 lb per hour. No published test of the Ultra exists.
Unknown
Unverified
GoveeLife Smart Nugget Pro
60 lb / 24 h, 2.5 lb per hour
Reviewed’s Leigh Harrington timed a full 3.5 lb bucket in “just over 90 minutes” — 3.5 ÷ 1.5 = 2.33 lb per hour, or 56 lb across 24 hours
−7%
Broadly holds
Frigidaire EFIC256
44 lb / 24 h
Reviewed’s Jonathan Chan measured “about 24 ounces of ice in an hour” — 1.5 lb per hour, or 36 lb across 24 hours
−18%
Overstated
Two things follow from that table, and the second one is the reason it leads this page.
First, the Frigidaire’s 44-pound badge is 18 percent optimistic. Chan’s figure is a sustained hourly rate, and no countertop machine holds its first-hour pace all day, so treat 36 as the ceiling and not the floor.
Is the GE Opal 2.0 Ultra’s Claim Verified?
The $449 machine’s number is a blank, and that is worse than the Frigidaire’s overstatement. Reviewed’s Opal write-up is genuinely useful, but it covers the Opal 2.0 rated at 24 pounds a day. The Ultra is a different SKU with a different rating. Nobody has published a measurement of it. Every “38 pounds a day” you will read today, including on the retail listings, traces back to GE.
A tall glass packed with small pellet ice.
Has Noise on Any Nugget Ice Maker Been Measured?
Nobody has metered the noise of any of these three, though you will see it quoted everywhere as though it were data. GoveeLife publishes 40 dB. Reviewed calls the Frigidaire “fairly noisy” with “random growling noises,” calls the Opal quiet, and puts the GoveeLife between them. Those are careful adjectives from people who listened to the machines, not decibels. If a page gives you a dB figure for the Opal or the Frigidaire, it made it up or copied a spec sheet.
Nugget Ice Maker Running Costs: Filters, Descaler and Power
The purchase price is the smaller half of this decision for exactly one of the three. GE’s parts store sells the scale-inhibiting filter the Ultra uses, part P4INKSFLTR, at $29 each, and GE’s own listing for the replacement pack says to replace it every three months. Four a year is $116. Add the six-pack of descaling powder at $15 and the Opal costs $131 a year to run before you plug it in.
The Frigidaire costs nothing. Its manual says the unit has no separate water filter; cleaning is diluted detergent, warm water and a cloth, and fresh reservoir water every 24 hours. That is more work and zero dollars.
The GoveeLife sits between them and I cannot price it. It needs no water filter, its cleaning reminder fires every 100 operating hours, and it runs a 10-minute quick clean or a 30-minute descale on its own. Govee sells cleaning and descaling agents as an accessory. Neither Govee’s US storefront nor any source I could reach lists a price for them, so the figure below is marked unknown rather than guessed.
Electricity is a smaller number and a shakier one. GE does not publish a wattage for the Ultra; the plate says 120 V and 2.5 A, which is a 300 W ceiling, not a draw. Govee’s peak is 210 W. The Frigidaire’s rating label reads 185 W. At the June 2026 US residential average of 18.34 cents per kWh from the EIA, four hours of run time a day at those rated ceilings works out to $80 a year for the Opal, $56 for the GoveeLife and $50 for the Frigidaire. Read those as upper bounds. No one has metered any of these machines, and a compressor cycling on a thermostat does not sit at its rating.
Which Nugget Ice Maker Scores Highest on the Tally?
Five axes, weighted for a machine that lives on a kitchen counter and gets used most days. Every weight is a judgment and every one is written down, so you can move them and re-run the tally yourself.
Verified output, 25%. The heaviest weight, because it is the single spec people buy on and the only one an outside party has ever checked.
Running cost, 20%. A $131 annual filter habit is a third of a Frigidaire every year.
Upkeep burden, 20%. Every one of these needs descaling. What differs is whether the machine reminds you and does it itself.
Warranty and maker backing, 20%. These are pumps and compressors in a wet box. Coverage length is not a footnote.
Purchase price, 15%. The lightest weight on purpose. A $167 price gap is real but recoverable; a machine that quits in month 14 is not.
Axis (weight)
GE Opal 2.0 Ultra
GoveeLife Nugget Pro
Frigidaire EFIC256
Verified output (25%)
4 — no published test of this SKU
8 — timed at 93% of claim
5 — measured, 18% short
Running cost (20%)
3 — $131/yr
7 — nothing required, price of optional agents unpublished
10 — $0
Upkeep burden (20%)
5 — filter and UV chute help, but GE publishes a known issue tying that filter to false Add Water faults
7 — self-clean cycles and a 100-hour reminder
4 — manual descaling, daily water change, no reminders
Warranty and backing (20%)
4 — 1 year
9 — 3 years, whole machine
3 — 1 year, and one Home Depot owner reports paying return shipping under it
Purchase price (15%)
4 — $449
5 — $424.99
8 — $281.75
Weighted tally
40.0
73.5
58.5
How Is the Weighted Score Calculated?
Worked example, so the arithmetic is not hiding: the GoveeLife scores (8 × 0.25) + (7 × 0.20) + (7 × 0.20) + (9 × 0.20) + (5 × 0.15) = 2.00 + 1.40 + 1.40 + 1.80 + 0.75 = 7.35, or 73.5 out of 100.
Disagree with a weight? Move it. If output is all you care about and you drop warranty to 5 percent, the GoveeLife wins by more. If you weight polish and app control — neither of which is an axis here, because neither is measurable — the Opal climbs. What it cannot climb on is the evidence.
GE Profile Opal 2.0 Ultra: A Rating No Lab Has Checked
$449 direct from GE, $499 at Home Depot, $448 at Abt. It is the nicest object here. Wi-Fi scheduling, a UV light in the dispense chute, a scale-inhibiting filter, and a build that reads as an appliance rather than a gadget.
Two things undercut it. The filter that is supposed to reduce upkeep has its own GE support article, titled around seeing “Add Water” or not making ice when using the scale-inhibiting filter. When the maker publishes a known-issue page about its own headline feature, that is worth more than a hundred five-star reviews. And on Best Buy’s 120 reviews the Ultra averages 4.0 stars with 17 of them — 14 percent — at one star, which is a heavier bottom tail than the average suggests. The recurring themes there are noise during operation, the sanitize-every-25-hours and descale-every-100-hours cadence, and false Add Water faults.
Skip this if you would rather not spend $131 a year on filters and descaler to run a machine whose output figure has never left the manufacturer’s own lab.
GoveeLife Smart Nugget Ice Maker Pro: the only claim that held
$424.99 on Govee’s store, $499.99 list, $499.99 at Best Buy where it currently shows as unavailable from that seller. 60 pounds a day claimed, 56 measured. Six minutes to first ice, per Govee. A 2.2-litre main tank plus a 4.9-litre side tank. Three-year warranty on the whole machine, which is the outlier in a category where one year is normal.
Here is the honest weakness, and it is not on the spec sheet. This machine showed up around CES in January 2026. Best Buy has 59 reviews at 4.14 stars. Govee’s own product page shows zero ratings. Seven months on the market is not a reliability record, and I am not going to pretend the three-year warranty is the same thing as three years of owners reporting back. It is a promise, not a track record.
Skip this if you want a machine with years of owner failure data behind it, or if a 21-inch depth with the tray attached will not fit your counter run.
Frigidaire EFIC256: no filter, no app, no cushion if it quits
$281.75 at Home Depot. 44 pounds claimed, 36 measured. 3-pound basket, 3-quart reservoir, 185 W, 36 pounds on the counter. No filter to buy and none to fail.
A glass of ice with condensation forming on the outside.
Something buyers should know that the branding hides: Frigidaire does not build this. Frigidaire-branded countertop ice makers are made under licence by Curtis International Ltd., and curtisint.com is the manufacturer site that lists the EFIC255 and EFIC256 as current products. Reviewed raises quality-control concerns about that arrangement directly in its EFIC255 review. The ratings line up with the concern rather than against it: 3.2 stars from 6 reviews on the EFIC256 at Home Depot, 3.2 stars from 46 reviews on the previous EFIC235, and Reviewed reports the EFIC255 at 3.1 stars on Amazon. Three separate retailers, two product generations, the same low-three rating.
One caution on checking availability yourself: frigidaire.com’s storefront will not render product or stock information to anything but a full browser, so the live listing at Home Depot and the current-product listing at Curtis International are the two places I could actually confirm this model is still sold.
Skip this if the machine will live within earshot of a conversation, or if a 3.2-star average across two generations is the kind of signal you would rather not argue with.
Recalls, and the two brand names people keep confusing
No open recall covers any of these three machines. Two adjacent facts are worth having anyway, because both get mangled online.
Curtis International, the company that actually builds the Frigidaire-branded machine, is under an expanded CPSC recall for Frigidaire-brand minifridges, model EFMIS121, announced 15 January 2026 — roughly 330,000 units added to a 2024 recall of 634,000, over internal components that can short and ignite the housing, with at least six fire reports. The CPSC notice covers no EFIC ice maker. Same licensee, different product line. I am naming it because I recommend one of that licensee’s products below and you should be able to weigh it.
The other one is a name collision. The widely repeated “nugget ice maker recalled for laceration hazard” story is Gevi, model GIMN-1102, roughly 8,800 units, 263 reports of a metal blade detaching into the ice basket. Gevi is not Govee. Several aggregator sites have run them together. The GoveeLife machine here has no recall; Govee does have a 2025 CPSC recall, and it covers space heaters.
The Tie-Breaker: What I Would Buy at These Prices
The tally says GoveeLife. My money says Frigidaire, and the gap between those two sentences is the whole point of showing you the weights.
The GoveeLife wins the Ledger on a three-year warranty and a verified output claim. Both are real. But it is a seven-month-old product with 59 reviews, its own maker’s page carries no ratings at all, and Best Buy has it listed as unavailable. That is a lot of confidence to buy at $424.99.
Why Buy the Frigidaire Instead?
The Frigidaire EFIC256 costs $281.75, and its claim is overstated by 18 percent, but its 36 measured pounds a day is still more ice than the Opal claims. It needs no filters. If it dies in month 14, which its 3.2-star history says is a live possibility, you are out $282 instead of $449 — and the $167 you did not spend covers most of a replacement. That is the trade I would take.
The one I would not buy is the GE Profile Opal 2.0 Ultra. Highest price, shortest warranty, $131 a year in consumables, and an output number no independent lab has ever put on a scale. It may well be the nicest machine here. Nobody outside GE has measured it.
Prices are approximate, current as of publication, and change frequently — confirm the latest price with the retailer before buying.
One structural note that applies to all three and gets left out of most comparisons: none of these bins is a freezer. Ice melts back into the reservoir and gets remade, so daily output only matters until the basket fills, and no basket here holds four pounds. If what you actually want is a cold drink that stays cold for eight hours, an insulated bottle answers that for a fortieth of the money. If the ice is for a drinks station, the blender comparison covers the frozen half and the cooler comparison covers the overflow. The rest is in Kitchen & Cooking.
If you meter one of these three with an actual sound level meter or a Kill A Watt, send me the numbers. The contact page works.
Some links on our site may earn us a commission at no extra cost to you. Learn more.
Water-safety note: whatever board you buy, wear a leash and a properly fitted personal flotation device, follow U.S. Coast Guard requirements for your waterway, and check local conditions before paddling. Follow the manufacturer’s inflation and weight-capacity guidance.
The inflatable vs hard paddle board decision comes down to one honest question: where will the board live when you’re not paddling? For most people — anyone with a car instead of a truck, a closet instead of a garage — an inflatable is the right answer, and modern drop-stitch inflatables are stable, durable, and shockingly capable. Hard boards still win on glide and responsiveness, but they demand roof racks and wall space that most buyers don’t have.
Our picks at a glance
Best overall: Bote Wulf Aero — OutdoorGearLab’s Editors’ Choice; a 9.0 stability score with easy handling for any skill level.
Best budget: Advenor Extra Wide — around $220 with a perfect 10 stability score and a 330 lb capacity; the beginner’s bargain.
Best upgrade: iROCKER Blackfin Model V — 12’6″ of length and a huge 485 lb capacity for paddlers who bring a dog, a kid, or a cooler.
Best for touring: Red Paddle Co Voyager+ — premium construction and the glide to cover real distance.
Inflatable vs. hard: the real trade-offs
Attribute
Inflatable SUP
Hard (epoxy) SUP
Glide & speed
Good; slight flex costs efficiency
Best; rigid hull tracks and carries speed
Stability for beginners
Excellent — thicker, wider shapes
Good, varies by shape
Storage
Duffel-bag sized when deflated
Needs rack or garage wall
Transport
Any car trunk
Roof rack or truck bed
Durability
Bounces off rocks and docks
Dings and cracks require repair
Setup time
8–12 minutes of pumping
Carry it to the water
Typical price range
~$220–$1,000+
~$700–$2,000+
The one hard-board advantage that matters daily is the absence of setup: you carry it in and paddle. If you live on the water, that convenience compounds and a hard board deserves a look. For everyone else, ten minutes of pumping buys you a board that fits in a hatchback and survives rental-fleet abuse — which is why every board on our shortlist below is an inflatable.
Ten minutes of pumping buys a board that packs into a trunk and shrugs off dock scrapes.
What you pay per pound of capacity
Capacity is the spec that quietly decides whether a board works for you — it has to float you, your gear, your dog, and the drinks. So we computed something board makers never print: dollars per pound of rated capacity, using OutdoorGearLab’s tested lineup and listed prices.
Board
Price
Size
Capacity
$ per lb of capacity
Advenor Extra Wide
$220
11′ × 33″
330 lbs
$0.67
FunWater 11
$231
11′ × 33″
330 lbs
$0.70
iROCKER Blackfin Model V
$999
12’6″ × 32″
485 lbs
$2.06
Bote Wulf Aero
$579
10’4″ × 34″
250 lbs
$2.32
Isle Pioneer 3
$695
10’6″ × 34″
285 lbs
$2.44
Red Paddle Voyager+
$1,749
13’2″ × 30″
330 lbs
$5.30
The spread is nearly 8x. Budget boards deliver capacity astonishingly cheaply; what the premium boards sell is construction stiffness, hardware quality, and glide — real things, but not things a first-summer paddler can feel yet. Our advice falls out of the table: start under $250, and upgrade only once you know you paddle often enough to notice the difference.
Which paddle board for which paddler?
First board, calm lakes, casual summers
The Advenor Extra Wide ($220) earned OutdoorGearLab’s Best Buy award with a perfect 10 stability score — at 33 inches wide it’s a floating dock you can learn on, and its 330 lb capacity handles an adult plus a kid or cooler. The FunWater 11 ($231) is the same idea with what testers called the best blend of “glide, stability, price, and fun” for beginners. Either is the right first board; flip a coin on price.
A wide, stable deck is exactly what makes a calm-lake first board feel effortless.
The everyday board you’ll keep for years
The Bote Wulf Aero ($579) took the Editors’ Choice award for a reason: a 9.0 stability score and a 7.5 glide score is a rare combination — steady enough for a nervous beginner, lively enough that you won’t outgrow it in a season. At 16.5 lbs it’s also one of the lightest here, which matters every time you carry it from the car. Its one real limit is the 250 lb capacity: bigger paddlers or dog-plus-gear loads should look at the Isle Pioneer 3 ($695, 285 lbs) — a similarly friendly all-arounder with more float.
Big loads, big paddlers, passengers with paws
The iROCKER Blackfin Model V ($999) is the truck of the category: 12’6″ long with a 485 lb capacity — the highest in OutdoorGearLab’s test — and deck rigging for serious gear. It weighs 28 lbs and costs real money, but nothing else here carries two people, a dog, and a weekend’s worth of stuff without sitting low in the water.
Distance days and expedition ambitions
The Red Paddle Co Voyager+ ($1,749) is the specialist: 13’2″ long and just 30 inches wide, built for covering miles with gear. Its price-per-capacity number is the worst on our table because you’re not buying capacity — you’re buying hull efficiency and Red Paddle’s famously stiff construction. If “paddle to a campsite” is the dream, this is the board; if it isn’t, save $1,100.
How we compared
Board specs, prices, and stability/glide scores come from OutdoorGearLab’s published side-by-side testing of 2026 paddle boards; the price-per-pound-of-capacity analysis is our own calculation from those figures. We cross-checked category trade-offs against long-term owner feedback on storage, pumping effort, and durability. We have not paddled these boards ourselves and don’t claim to — see our editorial standards for how we work. Prices are approximate, current as of publication, and change frequently — confirm the latest price with the retailer before buying.
Frequently asked questions
Are inflatable paddle boards worth it compared to hard boards?
For most recreational paddlers, yes. Modern drop-stitch inflatables are more stable and more durable than entry hard boards, and they fit in a car trunk. Hard boards win for speed, surf, and paddlers with waterfront storage.
How long does it take to inflate a paddle board?
Plan on 8–12 minutes with the included hand pump to reach the manufacturer’s recommended PSI. A $130–$180 electric pump that runs off a car outlet turns that into zero effort — the most popular first upgrade owners make.
What size paddle board do I need?
For all-around use, most adults do well on a board 10’–11′ long and 32″–34″ wide, with a rated capacity at least 60–80 lbs above your body weight to account for gear, water conditions, and passengers.
Do I need a life jacket on a paddle board?
On most U.S. waters, yes — the Coast Guard classifies SUPs as vessels beyond swimming areas, which generally means carrying a PFD (and wearing it is the smart default). Check your state and local rules, and always use a leash.
Bottom line
Buy inflatable unless you have waterfront storage and a need for speed. Start with the Advenor or FunWater if this is your first summer on the water, step up to the Bote Wulf Aero when you know you’re hooked, and reserve the Blackfin V and Voyager+ for the specific jobs they’re built for. Rounding out a water-day kit? Our cooler face-off, beach shade comparison, and dry bag vs. waterproof backpack guide cover the rest, with more in Outdoors & Travel.
Some links on our site may earn us a commission at no extra cost to you. Learn more.
Updated August 2026: Mac Sports’ own page lists the MACWagon at 27.1 lb; Easterly publishes two load ratings, a 150 kg static figure and an 80 kg dynamic one, not just the single 330 lb figure most roundups repeat. The RTIC Ultra Tough is sold out at RTIC as of this update. New here: the October 2025 federal recall of a folding utility wagon under the stroller rule, and what each of these four says in writing about children riding.
The Veer All-Terrain Cruiser takes the tally at 79.9 out of 100 and the RTIC Ultra Tough finishes last at 39.7. That order flips the moment you decide the wagon is only ever hauling cargo: strike the passenger axis and the $169.99 Mac Sports MACWagon wins at 87.5 with the $699 Veer second. Which of those two beach wagon tallies is yours depends on one question, and it is not about sand. It is whether a child is going to sit in it.
Weigh These Five Things, Not the Beach Wagon’s Sticker
Every score below is computed from a figure the maker publishes on its own site, or from a document a government agency publishes. Where a maker publishes nothing, the cell is a zero with the reason printed rather than a guess.
Rolling capacity — 25%. The lowest load figure the maker will stand behind for a wagon in motion, divided by the highest in this field, 300 lb. Not the parked figure.
Passengers permitted and certified — 20%. Does the maker allow a person to ride, and is the wagon certified to a child-carrying standard? 100 if yes, 0 if the maker says no or says nothing.
Price per pound of rolling capacity — 20%. The maker’s own current price divided by the rolling figure above, then scored against the cheapest.
Buyable today on the maker’s site — 15%. 100 for an add-to-cart button, 0 for a notify-me form.
Warranty published with readable terms — 20%. Months of coverage divided by 24, the longest verifiable term here. A warranty you cannot find scores zero.
Wagon
Capacity 25%
Passengers 20%
Price/lb 20%
Buyable 15%
Warranty 20%
Total
Veer All-Terrain Cruiser
83.3
100
20.3
100
100.0
79.9
Mac Sports MACWagon Push & Pull
100.0
0
100.0
100
50.0
70.0
Easterly Pull-About
58.8
0
25.0
100
50.0
44.7
RTIC Ultra Tough Wagon
83.3
0
94.4
0
0
39.7
The Veer’s 79.9 is (83.3 × 0.25) + (100 × 0.20) + (20.3 × 0.20) + (100 × 0.15) + (100.0 × 0.20) = 20.83 + 20.00 + 4.06 + 15.00 + 20.00. It wins on one axis worth a fifth of the tally and it is the worst value per pound in the field by a factor of nearly five. That is a fragile win, and the section near the bottom shows exactly how fragile.
Read the Capacity the Maker Stands Behind While Rolling
A rated capacity on a folding wagon is not one number. Easterly makes that plain on its own product page and everyone quoting it drops half the answer.
Claim
Where it is printed
What the maker’s own page also says
Verdict
Easterly Pull-About holds 330 lb
Field Mag’s wagon test, 17 June 2026
easterly.co prints a static load of 150 kg AND a dynamic load of 80 kg — 330.7 lb parked, 176.4 lb moving
The headline figure is the parked figure
MACWagon weighs 36.6 lb empty
Field Mag’s spec table
macsports.com prints a product weight of 27.1 lb
9.5 lb apart. I could not resolve it and did not pick one
Veer Cruiser holds 250 lb
goveer.com
The same page prints 55 lb per seat across 2 seats for passengers — a separate limit for a separate job
Two limits, both published, neither hidden
RTIC Ultra Tough is ‘built to carry up to 250 lbs’
rticoutdoors.com
The same page publishes no empty weight, no wheel size and no warranty
Nothing to check the claim against
Static and dynamic are not marketing hedges. A static load is what a structure holds standing still. A dynamic load is what it holds while the thing is rolling, dropping off a curb, or being yanked over a rut. Easterly rates its wagon at 150 kg one way and 80 kg the other, and the second number is 53 percent of the first. Every roundup I read, including Field Mag’s editor-tested folding wagon comparison, prints 330 lb and stops.
So the tally above uses 176.4 lb for the Easterly. That is Easterly’s own number for a wagon in use, and a beach wagon is never parked when it matters.
A loaded family setup is chairs, a cooler, an umbrella and a bag. The rating that matters is the one for a wagon in motion.
Look Up the Stroller Rule Before You Put a Child In One
Three of these makers print a warning. Mac Sports writes on the MACWagon page: ‘Please note that it is not made to transport children.’ Easterly writes on the Pull-About page: ‘The Pull-About is not designed for transporting children, adults or pets.’ Target’s Embark Collapsible Wagon, the $50 answer people reach for first, says it is ‘not recommend for carrying children’ and was out of stock when I checked. RTIC’s product page says nothing either way, which is not the same as saying yes.
The reason those sentences exist is regulatory. 16 CFR Part 1227 requires that each carriage and stroller comply with ASTM F833-21, and CPSC’s own business guidance defines a stroller as ‘a wheeled vehicle for the transport of one or more infants or children generally in a sitting-up or semi-reclined position.’ That guidance does not use the word wagon anywhere. A wagon a child sits up in still fits the definition.
And the agency has enforced it that way. On 23 October 2025 CPSC published recall 26-044: Olympia Tools International recalled about 530 Pack-N-Stroll Premium Folding Utility Wagons, sold through Costco.com and cohooutdoors.com from June through July 2025 for $200 to $400. The stated reason is that the wagon ‘violates the mandatory standard for strollers because an opening between the organizational tray and the sidewall may cause a child’s head to become entrapped, posing a risk of serious injury or death,’ and that ‘the utility wagon’s lack of a restraint system poses a risk of falls to children.’ No incidents were reported. The remedy is a refund.
Read that hazard sentence again. Nothing broke. The wagon was recalled for what it does not have: a gap small enough and a harness at all. Those are exactly the things a cargo wagon has no reason to include.
One of these four is built the other way. Veer publishes a passenger rating of 55 lb per seat across two seats, states that the Cruiser is JPMA Certified, and claims it ‘meets or exceeds ASTM, SOR and EN 1888.2 safety standards.’ That is a different product class wearing a wagon’s shape, and it is priced like one at $699.
This page reports what each maker and the CPSC publish. It does not rank these wagons by safety and it is not a substitute for the manufacturer’s own instructions. If a child is going to ride, follow that maker’s harness, age and weight guidance exactly, and do not put a child in a wagon whose maker says not to.
I searched the CPSC recall record for all four brands and found nothing covering the MACWagon, the Cruiser, the Ultra Tough or the Pull-About.
Divide the Price by the Pounds It Will Actually Roll
Now the money, computed on the rolling figures rather than the parked ones. Prices are each maker’s own, read this month.
Wagon
Maker’s price
Rolling capacity used
Dollars per pound
Mac Sports MACWagon Push & Pull
$169.99
300 lb
$0.57
RTIC Ultra Tough Wagon
$149.99 (sold out; list $249.99)
250 lb
$0.60
Easterly Pull-About
$399.00
176.4 lb (80 kg dynamic)
$2.26
Veer All-Terrain Cruiser
$699.00
250 lb cargo
$2.80
The spread from top to bottom is 4.9 times. Two things are worth pulling out of that table rather than leaving in it.
First, the RTIC’s $0.60 is a sale number. At its own list price of $249.99 it is $1.00 a pound and its score on that axis falls from 94.4 to 56.7, which would drop it further behind the Easterly. Second, the Easterly moves from $1.21 a pound to $2.26 depending purely on which of its two published ratings you use. That is the largest single swing on this page and it comes from reading one more line on the maker’s own site.
Buy the MACWagon If the Load Is Cargo
At $169.99 with a 300 lb rating, the MACWagon is the cheapest pound-for-pound wagon here and the only one that took a perfect score on two axes. Field Mag’s testers put it at the top on the surface this page is named for: ‘The MacWagon’s wheels made it the easiest to move through the obstacle course being both the heartiest on sand and also moved the most fluid on concrete.’ That is the closest thing to a sand verdict anyone has published on this field.
Two honest problems. Mac Sports covers it for one year, half of what Veer publishes. And the empty weight is genuinely unsettled: Mac Sports prints 27.1 lb on its own page while Field Mag’s table prints 36.6 lb. I am not picking between them. Note that Field Mag’s table also prints 36.6 lb for the Ozark Trail cart on the same page, which is the kind of coincidence that makes a spec table worth doubting rather than quoting.
Skip this if anyone is going to ride in it. Mac Sports has written down that this is not what the wagon is for.
Pay $699 Only If Children Ride
The Veer All-Terrain Cruiser wins this tally and it wins it narrowly and for one reason. Take the passenger axis away and it drops to second. On cargo alone it costs 4.9 times what the MACWagon costs per pound it will carry, and it carries 50 lb less.
What the money buys is the thing the other three decline to sell you: a published passenger rating, a restraint system, and JPMA certification. Veer also has the longest published warranty here, two years, and its own warranty page states it runs ‘from your date of purchase’ rather than from registration. At 32.6 lb with wheels on and 24.6 lb with them off, it is the only one that gets meaningfully lighter for the trunk.
Skip this if the only passengers are a cooler and four chairs. You are paying $529 more than the MACWagon for a capability you will not use, and getting less capacity for it.
Read Easterly’s Two Load Numbers Before Paying $399
Field Mag named the Pull-About its best overall wagon. My tally puts it third at 44.7, and that disagreement is worth stating plainly rather than burying: it is not a judgment about how the thing rolls. It is about what happens when you score it on the rating Easterly publishes for a wagon in motion instead of the one everybody quotes.
Credit where it is due. Easterly is the only maker in this field that publishes both a static and a dynamic figure at all. The other three publish one number and leave you to assume which condition it describes. Easterly told the truth in more detail and got scored down for it, which is a real weakness in how I have weighted this and you should know it before you accept the order.
The rest: $399, in stock with an add-to-cart button, 27.6 lb empty, inflatable rear tyres with front casters, and a 12-month warranty from date of purchase per Easterly’s own warranty page. The passenger line is unambiguous and covers adults and pets as well as children.
Skip this if you were buying it on the 330 lb figure. At 176 lb rolling it is a mid-capacity wagon at a premium price.
Wait for the RTIC, or Move On
The named loser, at 39.7, and it lost on things RTIC could fix this afternoon. The Ultra Tough Wagon is sold out on rticoutdoors.com with a notify-me form where the buy button goes. Its product page publishes a capacity and nothing else — no empty weight, no wheel diameter, no warranty term. Field Mag lists it at 30 lb and $150 and rates it best for camping, praising ‘a thicker, heartier, durable steel frame’ over its nearest rival.
It is probably a good wagon. It scored 94.4 on price per pound, second only to the MACWagon. But two of five axes are zeros because RTIC has published nothing to fill them with, and a wagon you cannot buy is not a recommendation.
Skip this if you need a wagon before the weekend, or if you want to know what it weighs before it arrives.
Zero the Passenger Axis and Watch the Order Change
Most people buying a beach wagon are hauling gear. For them the passenger axis is worth nothing, so here is the same tally with it removed and its 20 points redistributed across the other four axes in proportion.
Wagon
Full tally
Cargo-only tally
Change
Mac Sports MACWagon Push & Pull
70.0
87.5
+17.5, and first place
Veer All-Terrain Cruiser
79.9
74.9
−5.0, second
Easterly Pull-About
44.7
55.9
+11.2, third
RTIC Ultra Tough Wagon
39.7
49.6
+9.9, fourth
One axis, a fifth of the weight, and the winner changes and the gap becomes 12.6 points instead of 9.9 the other way. If you want the arithmetic: with the passenger axis gone the remaining weights become 31.25, 25, 18.75 and 25 percent, and the MACWagon’s 87.5 is (100.0 × 0.3125) + (100.0 × 0.25) + (100 × 0.1875) + (50.0 × 0.25).
Prices are approximate, current as of publication, and change frequently — confirm the latest price with the retailer before buying.
Take Two Numbers Away From This
The first is 80 kg, because it is Easterly’s own rating for the wagon while it is moving and it is 53 percent of the figure printed everywhere else. Before you buy any wagon on a capacity claim, find out whether that claim describes it parked or rolling. Three of these four makers do not say.
The second is recall 26-044, because it settles an argument people have every summer. A folding utility wagon was pulled off the market under the federal stroller standard for a gap and a missing harness, and three of the four wagons on this page carry a written instruction from their own maker not to put a child in them.
To provide the best experiences, we use technologies like cookies to store and/or access device information. Consenting to these technologies will allow us to process data such as browsing behavior or unique IDs on this site. Not consenting or withdrawing consent, may adversely affect certain features and functions.
Functional
Always active
The technical storage or access is strictly necessary for the legitimate purpose of enabling the use of a specific service explicitly requested by the subscriber or user, or for the sole purpose of carrying out the transmission of a communication over an electronic communications network.
Preferences
The technical storage or access is necessary for the legitimate purpose of storing preferences that are not requested by the subscriber or user.
Statistics
The technical storage or access that is used exclusively for statistical purposes.The technical storage or access that is used exclusively for anonymous statistical purposes. Without a subpoena, voluntary compliance on the part of your Internet Service Provider, or additional records from a third party, information stored or retrieved for this purpose alone cannot usually be used to identify you.
Marketing
The technical storage or access is required to create user profiles to send advertising, or to track the user on a website or across several websites for similar marketing purposes.
