Ranking Every Subaru Boxer Engine From GRENADE to BULLETPROOF

One Subaru engine has ended more Craigslist deals, more forum arguments, and honestly more actual friendships than any other single part this brand has ever built. And it's probably not the one you're thinking of right now. Everybody assumes Subaru reliability is one flat answer. Either you trust the brand, or you don't. That's completely wrong. And it's the single Subaru has built more than a dozen distinct boxer engines under that same flat-four badge since 1989, spread across compacts, wagons, rally cars, and sports coupes. And the gap between the worst one and the best one isn't a small gap. It's the difference between an engine that can grenade itself before it clears its own factory warranty, and an engine so overbuilt that mechanics half-jokingly call it unkillable. By the end of this video, you're going to know exactly where your engine falls on that ladder. And if you own one of the two I'm putting at the very bottom, I need you to actually stick around. Because what I'm about to tell you might save you thousands of dollars before your next oil change. Here's the thing almost nobody explains properly. The boxer layout itself, pistons punching sideways at each other instead of standing up and down like a normal engine, is not the problem and never has been. It's a genuinely brilliant piece of engineering, low center of gravity, natural balance, the whole reason Subarus feel planted in a way front-engine competitors don't. The problem has never been the layout. It's been specific material choices, specific piston ring designs, specific gasket compounds, made by specific engineering teams at specific points across three and a half decades. And every single one of those decisions produced a completely different reliability outcome, wearing the exact same boxer name on the valve cover. Lump them all together, and you're doing exactly what gets people burned on a used car lot. Separate them out, and suddenly the whole picture makes sense. I spent a chunk of my 20s turning wrenches at an independent shop that did nothing but Subaru and Toyota work. And I want to tell you about the day that actually taught me this lesson. Because it wasn't a lecture, it was a tow truck. A guy rolled in towing a Forester with barely 60,000 miles on it, engine completely seized, oil light apparently ignored for who knows how long, and he was furious, convinced Subaru had sold him a lemon and that the whole brand was a scam. Same week, same shop, a rally guy dropped off a WRX with 280,000 miles on the original engine, still running strong, tired clutch and not much else. There for routine work he'd been putting off. Same badge on both hoods, same general boxer layout, completely different engines under the sheet metal, and completely different outcomes waiting for two guys who both thought they understood what Subaru reliability meant. That gap is the entire subject of this video, and it starts at the absolute bottom of the ladder, with the engine that earned the FB25 Phase 1, built roughly 2011 through 2013. If you've watched anything else on this channel about Foresters or Outbacks from this window, you already know part of this story. But here's the engine-specific version. Subaru introduced the FB25 to replace the aging EJ25, and on paper it was supposed to be the fix. A cleaner running, more efficient successor. Instead, engineers specced low-tension piston rings into that first production run, specifically to reduce internal friction and squeeze out better EPA fuel economy numbers on the window sticker. Those low-tension rings don't seal against the cylinder wall the way a standard ring does, and in practice, that let oil slip past them into the combustion chamber and burn away with the fuel, instead of staying in the crankcase where it belonged. Owners reported losing a quart of oil every 1 to 2,000 miles, with zero visible leak anywhere underneath the car. Which is exactly the kind of failure that fools even careful owners, because there's nothing to actually go looking for, and the number of engines this actually killed is worse than most people assume. Enough of these engines hit total seizure, some with barely 50,000 to 70,000 miles on the odometer, that Subaru ended up settling a class action lawsuit over it, quietly replacing short blocks under extended warranty programs most owners never even knew existed unless they specifically called Subaru Corporate and asked the right question. This is the one engine on this entire list that earns the grenade label without any argument, without any tribal forum debate defending it. If you've got a 2011 through 2013 Forester, Outback, Impreza, or Crosstrek with the FB25, checking your oil level isn't optional maintenance anymore. It's the difference between catching this early and financing a new engine, and the engine directly above it on this ladder is the one everyone actually blames for Subaru's head gasket reputation, even though it's a completely different failure story. One tier up from rock bottom, and this one's going to make some of you defensive, the EJ25 Soaks, the naturally aspirated 2.5 liter that powered Foresters, Outbacks, Legacies, and Imprezas from roughly 1999 through 2010. This is the engine most people actually mean when they complain about Subaru head gaskets, and the mechanism here is different from the FB25's oil consumption problem entirely. The composite head gasket material Subaru used in this engine simply wasn't built for the internal pressures the boxer layout generates over a decade of heat cycling. Every time the engine warms up and cools down, the aluminum heads expand and contract at a different rate than the block underneath them, and that gasket has to absorb that constant flexing for years without ever fully sealing. It's a design compromise, not a manufacturing defect in the traditional sense. Every single one of these engines built to this original spec carries the same underlying risk, which is exactly why the failure rate on this engine is so consistently documented across owner forums, insurance data, and independent shop records alike. But that's not even the part that should worry you most if you own one of these right now. Somewhere between 100,000 and 150,000 miles, that gasket starts to weep, coolant mixes into the oil, and you get a milky sludge that eats bearings from the inside. A proper repair runs $2,000 to $2,800 done right. And here's the part that makes people furious when they find out. A meaningful number of these gaskets have already been replaced once by a previous owner using the same flawed original spec material, which means the fresh repair on the Carfax can blow again within another 30,000 miles. So if you thought a documented head gasket replacement made a used Subaru safe to buy, you were only half right. And the next engine on this ladder is where the arguments in the comments section are about to get genuinely heated. Now here's where things get genuinely tribal. And I want you to actually comment and tell me if you disagree, because I guarantee half of you will. The EJ257, the 2.5 liter turbocharged engine that powered the STI from 2004 all the way through its final year in 2021. In stock form, driven sensibly, this engine is genuinely tough. It's the engine that built Subaru's entire rally and motorsport reputation. Factory rated around 305 horsepower depending on the year, and plenty of them cross 200,000 miles without drama. But push it past stock power levels, which an enormous percentage of STI owners do, because that's the entire point of buying an STI. And you run straight into the engine's most famous weakness, ringland failure. And it's one of the most argued about mechanical failures in the entire Subaru community. The factory pistons in this engine have a known weak point at the ringlands. The small ridges that hold the piston rings in place, and under sustained high boost, or an aggressive tune, especially one that runs lean under load, those ringlands can literally crack and fail. Sometimes catastrophically, sometimes with startling suddenness on a car that felt completely fine the drive before. A full rebuild after a ringland failure, done properly with upgraded forged pistons to actually fix the underlying weakness, instead of just replacing like for like, can run anywhere from $4,000 to $7,000 depending on how much of the bottom end needs attention. And that number alone is enough to start the exact fight in the comments right now. This is exactly the kind of engine where you'll find two STI owners in the same forum thread absolutely convinced the other one is wrong. One camp says the EJ257 is bulletproof, and anyone who's blown one just didn't know how to tune a car properly. The other camp says it's a ticking time bomb the second you touch the boost. The honest answer sits in the middle. Stock, maintained, driven like a normal car, it's a genuinely solid engine. Modified without real expertise behind the tune, it's one of the more failure-prone engines on this entire list. And that duality is exactly why it lands in the middle of our ladder instead of clearly at the top or bottom. One step further up, and we hit an engine that gets unfairly lumped in with its bigger STI sibling, the EJ205, the 2.0 liter turbo that powered the early WRX through the mid-2000s. It shares a family resemblance with the 257, including some of that same ringland vulnerability under heavy modification, but the smaller displacement and generally lower stock boost levels mean it tends to survive spirited daily driving better than people assume. The more common failure point on these isn't the bottom end at all. It's the Avxo's variable valve timing solenoids, and the factory turbo's relatively short service life, compared to later designs, both of which are wear items rather than design flaws, and both of which are far cheaper to address than a bottom-end rebuild. Owners who stay on top of turbo maintenance, change their oil on schedule with a quality synthetic, and don't chase huge power numbers regularly see these engines clear 200,000 miles, which puts this one solidly in the respect it, don't abuse it category, rather than anywhere near grenade territory. It's proof that displacement and boost pressure matter just as as the underlying architecture when you're figuring out how hard an engine can actually be pushed, and it sets up the exact opposite problem with the engine we're covering next, one that gets almost no attention at all, good or bad. Let's talk about the engine that almost never gets mentioned in these conversations, and honestly, that's the whole problem. The FB20, the smaller 2.0 liter non-turbo, that's been quietly doing unglamorous work in Impreza's and Crosstrek's for over a decade, no lawsuits, no forum wars, no viral failure videos, just a smaller displacement version of the FB architecture that dodged the low-tension ring issue that plagued its 2.5 liter FB25 sibling, because Subaru's fix for that oil consumption problem had already been implemented by the time most FB20 applications hit the market in volume. It's not exciting, it's not going to win a spec sheet argument against a turbo-charged anything, but it is, quietly, one of the more underrated engines Subaru has built in the last 15 years, and that's exactly the kind of engine that never gets its own video, right up until this one, right above it, the FB25 Phase 2, built from 2014 onward, and this is where Subaru actually redeemed themselves after the disaster we started this video with. Same displacement, same general architecture, as the engine sitting at the very bottom of our list, but with the piston ring design corrected, and years of real-world refinement folded in. If you're looking at a Forester or Outback with an FB25 built in 2014 or later, you're looking at a genuinely different engine from a reliability standpoint than the one that generated a class action lawsuit, even though the badge on the valve cover reads identical. This is proof of exactly the pattern this whole video is built around. The year and the specific engine revision matter infinitely more than the family name stamped on the block, which means there's a real chance you've been worried about the wrong engine this entire time. If you've made it this far, drop a comment telling me your engine code, not just your model, the actual code stamped on your engine bay sticker or block, because I want to see how many of you have been driving around not knowing which tier you're actually in. Climbing higher now, into the engines that get genuine enthusiast respect without much argument, the FA20, the naturally aspirated 2.0-liter that powered the BRZ and its Toyota badge twins. This one was co-developed with Toyota, built off the FB20 block, but with new pistons, rods, and a new crankshaft, plus Toyota's direct injection system layered in. It picked up a reputation early on for high oil consumption in certain driving conditions, and Subaru did issue service updates to address it, but structurally, this is a well-regarded engine. Ward's Auto World named it one of their top engines the year it launched, and mechanics who've spent real time with it generally back that reputation up. Driven sensibly and maintained on schedule, it's one of the more dependable naturally aspirated engines currently wearing the boxer name. Its turbocharged cousin, on the other hand, is about to become the second fight I promised you at the start of this video. Now here's the second tribal argument I promised you, and this one might get you even angrier than the STI debate. The FA20 DIT, the direct injection turbo that's powered the WRX since 2015. This engine brought real modernization to the platform, a timing chain instead of the EJ's timing belt, direct injection, genuinely impressive factory power for its size. And in stock form, it's a strong, dependable engine, but like its EJ257 predecessor, it inherited a very similar Achilles heel, ringland failure under aggressive tuning, particularly on setups running E85 with a tune that isn't dialed in correctly, where lean conditions under boost can crack those same vulnerable ringlands. The forums are, predictably, split exactly down the middle. Stock WRX owners will swear this engine is bulletproof. Owners who've blown one on a bad tune will tell you it's fragile. Both of them are describing the same engine under completely different load conditions. And that's the whole point. This isn't a manufacturing defect story like the FB25. It's a story about what happens when you take a genuinely solid factory engine and push it past what the factory ever designed it to survive. Now we're getting into genuinely elite territory, and this next one might surprise you if you've been assuming turbo always means fragile. The FA24F, the 2.4 liter turbo currently powering the Ascent, Outback XT, Legacy XT, and the newest generation WRX. This is Subaru's most recent turbo architecture, and by every account from the tuning and enthusiast community, it addresses the weak points that made the EJ257 and FA27DIT into forum battlegrounds in the first place. Independent tuners who've spent real time on this platform report meaningfully higher reliability margins, even at elevated power levels, compared to a built EJ, which for an engine family with this much modification history is a genuinely significant claim. It's still young enough that we don't have 15 years of real world data the way we do on the EJ family, but everything currently on the table points to Subaru actually learning from three decades of ring land complaints and building accordingly. It's not the only recent engine that benefited from that lesson either. It's naturally aspirated sibling, the FA24D, that debuted in the 2022 BRZ and Toyota GR86, deserves its own mention here too. It directly addressed two specific complaints that dogged the earlier FA20, a torque dip in the mid-range and high oil temperatures under sustained hard driving, and by most enthusiast accounts, it's simply a more mature, more sorted version of an engine family that was already respected. This is what it looks like when an automaker actually incorporates a decade of owner feedback into the next revision, instead of just changing the badge and hoping nobody notices. Before we get to the very top of this ladder, there's one more engine worth flagging, even though I can't fully rank it yet, and that honesty matters more than pretending I have data I don't, the CB18, Subaru's newest turbocharged boxer, currently running in the overseas market Levorg and Forester lineups. Early tuning shops that have gotten their hands on this platform report solid reliability in the 300 to 350 horsepower range on stock internals, which if it holds up over the next several years, would put it in genuinely elite company. But this engine simply doesn't have the mileage history yet to sit confidently anywhere on this ladder. I'm mentioning it here specifically, so you know it exists, and so you don't mistake silence for a verdict. Give this engine another five years of real-world abuse before you trust anyone's tier list on it, including mine. And now, at the very top of this ladder, the engine that started this entire boxer legacy, and still, decades later, gets mentioned in the same breath as bulletproof by mechanics who have genuinely no reason to be sentimental about it, the EJ22. This engine powered legacies and early outbacks through the 1990s, and it did not carry the same head gasket weaknesses that would come to define its EJ25 successor. Chain-driven variants exist with essentially nothing to time out over the engine's life, and independent Subaru specialists will tell you, without hesitation, that this is one of the most stubbornly reliable engines this brand has ever produced. The kind of engine that shows up in junkyards still turning over at 300,000 plus miles, not because nobody tried to kill it, but because nobody managed to. It's not fast, it's not exciting, it will never headline a tuning forum, but if bulletproof is the standard we're measuring against, this is the engine that set it. And honestly, everything Subaru has built since has been trying to recapture what this one did almost by accident. So here's how you actually use everything we just covered. First, find your engine code. It's stamped on a sticker in the engine bay, usually near the strut tower or on the block itself, and it takes 30 seconds to locate. Don't rely on model year alone, because as we've now covered in detail, the exact same model year can carry completely different engines depending on trim and region. Second, if you're sitting on an FB25 built in 2011 through 2013, start checking your oil level manually between service intervals today, not at your next scheduled oil change. That single habit is the cheapest insurance you can give yourself against the worst engine on this entire list. And third, if you're shopping for anything with an EJ25 Sox, budget mentally for a head gasket job somewhere in your ownership window, even on a car that looks perfect right now, because the timeline on that failure is measured in mileage, not luck. That handles the two engines most likely to cost you money without warning. But if you're anywhere near the modification side of this hobby, there are two more rules you genuinely cannot skip. Fourth, if you're modifying an EJ257 or an FA20DIT, the ringland failure stories you've heard aren't myths, and they aren't tuner incompetence propaganda either. They're a real, documented weak point. And the difference between an engine that survives your build and one that grenades on you usually comes down to whether your tuner actually understands how to keep that engine fed correctly under load, not how much power the dyno sheet claims you're making. And fifth, if you're buying used specifically because of something you heard in this video, don't just ask a private seller which engine is under the hood. Verify it yourself off the sticker, because plenty of sellers genuinely don't know the difference between an early FB25 and a later one, and an honest mistake on their end can still cost you thousands on yours. That guy who towed his seized Forester into our shop all those years ago ended up trading it for a legacy with the EJ22, on my recommendation, mostly as a joke about how simple and boring it would feel, after what he'd just been through. He put almost 200,000 miles on that thing, over the better part of a decade, drove it through two more house moves and a cross-country trip, and never once called the shop about the engine itself. That's the entire lesson of this video, condensed into one guy's driveway. The badge on the back of a Subaru tells you almost nothing on its own. The three or four characters stamped on the engine block tell you everything. Now you know exactly where every one of these engines actually falls, and if yours turned out to be closer to grenade than bulletproof, at least you know what to watch for before it decides to make that decision for you instead. If this helped you understand what's actually sitting under your hood, hit subscribe, because we break down exactly this kind of engine-level intelligence every single week, and I'll see you in the next one. Audit Summary. What Changed and Why? At 155 WPM, the original draft had 10 stretches at or over 60 seconds, with a flat, non-curiosity close. The FB25 mechanism beat. The FB25 lawsuit beat. Both EJ25 beats. The EJ257 mechanism beat. The worst offender at 89 seconds unbroken. The EJ205 beat. The FB25 Phase 2 beat. The FA20 beat. The FA24F beat. And, by far the biggest issue, the 5-point closing framework, which ran 117 seconds as a single unbroken block, with a flat list-style close and zero internal hook. Fixes applied. Every flagged beat now ends on a forward-pulling tease or number-slash callback, instead of a factual summary. E.g., the number of engines this actually killed is worse than most people assume. The exact fight I want you to have in the comments right now. There's a real chance you've been worried about the wrong engine this entire time. The 117-second framework got split into two beats around points 3 and 4, with a bridging line that handles the two engines most likely to cost you money without warning. But if you're anywhere near the modification side of this hobby, there are two more rules you genuinely cannot skip, so the 5-point list no longer runs as one unbroken stretch. The EJ257 mechanism beat got the same treatment, split before the Ringland failure explanation with its own microhook. Word count grew from Gubs3,400 to Gubs3,645, as a byproduct of adding genuine transitional content, not padding, which also pushes runtime further past the 20-minute floor.