8L90 vs 10L90: Differences, Interchange, and Which Is Stronger

Walk a GM rear-drive service bay and you will find these two units parked next to each other constantly. The 8L90 arrived for the 2015 model year and spent a decade behind small blocks in Silverados, Sierras, Escalades, Camaros and Corvettes. The 10L90 landed in the 2017 Camaro ZL1 and then spread into the performance and heavy-trim end of the same vehicle lines. Same customer, same driveway, two completely different transmissions.

The question that shows up on the phone is always some version of "which one is better," and the honest answer is that these two are not separated by torque capacity the way most people assume. They are separated by ratio spread, clutch count, fluid, and above all by the way each one fails. Here are the real numbers off GM's own documents.

Related reading: our 6L80 vs 8L90 comparison and 8L90 rebuild wear points.

---

Recommended 8L90 parts

In stock, ships same day from the USA. Secure checkout.

8L90 Master Rebuild Kit

$979.99View part

8L90 Pump 2015

$942.99View part

8L90 Valve Body 2015-2024

$860.99View part

The Numbers, Side by Side

The 8L90 column comes off GM Powered Solutions' published specification sheet for the unit (RPO AH2R, N8X ratio set). The 10L90 column comes off Chevrolet Performance's own Supermatic 10L90 service literature, part no. 19420481, which carries the factory ratio table, weight, torque rating and fluid capacities.

Read the torque rows before you read anything else. GM rates the 8L90 at 665 lb-ft maximum gearbox torque. Chevrolet Performance rates the crate 10L90 at 650 lb-ft. Whatever else the ten-speed brings, it is not a large jump in holding capacity, and it weighs about 19 lb more. Anybody selling a 10L90 swap as a strength upgrade over a healthy 8L90 is selling gear count.

---

What the Two Extra Gears Actually Bought

Both units run four planetary gearsets. The 8L90 gets eight forward ratios out of five clutches; the 10L90 gets ten out of six, adding a rotating clutch so it runs two brake clutches and four rotating clutches. GM put those four rotating clutches forward of the gearsets specifically to shorten the oil feeds, which is a large part of why the ten-speed's clutch-to-clutch shifts land as fast as they do.

The payoff shows up in the spread: 7.39 on the 10L90 against 7.0 on the 8L90, with a numerically steeper 4.696 first and three overdrives topping out at 0.636. In practice the ten-speed keeps the engine in a narrower rpm band through the whole pull and drops cruise rpm noticeably at highway speed. The 8L90's 7.0 spread with a 4.56 first and a 0.65 eighth is not far behind on paper, but it has to cover that range in two fewer steps, so each step is bigger and you feel it.

The pump is the other real difference

The 8L90 uses a conventional on-axis arrangement. The 10L90 moved to an off-axis variable vane pump that sits low in the unit, inside the valve body itself. Chevrolet Performance describes the design as effectively two pumps in the packaging size of one, which is how GM held parasitic loss down while feeding six clutches. It also means the pump, the valve body and the control hydraulics are one interconnected assembly, so pump vane wear becomes a valve body conversation instead of a front-of-the-case conversation.

---

Fluid: These Two Do Not Share a Bottle

This is the most common way a shop damages one of these units, and it is entirely avoidable. GM's specification sheet lists DEXRON High Performance ATF for the 8L90 at a 10.3 liter capacity excluding cooler volume. The 10L90 takes DEXRON ULV, and Chevrolet Performance's installation manual is blunt about it: use DEXRON ULV only (P/N 19352619 US, 19352620 Canada), and do not use DEXRON VI or DEXRON HP. The same manual notes something most people skip — the quart containers of ULV have to be shaken before pouring to stir the additive package back into suspension.

The 10L90 fill procedure, by the numbers

There is no dipstick. The level is set through a tube in the bottom pan, and the temperature window is narrow enough that getting it wrong ruins the job:

• Pan removal and filter replacement: 7.7 liters (8.1 quarts)
• Overhaul, transmission volume only: 10.95 liters (11.6 quarts)
• Complete system including cooler: 12.52 liters (13.23 quarts)
• Set the level with TFT between 75 and 80 C (167 to 176 F). Above 80 C you end up under-filled; below 75 C you end up over-filled.
• The unit has an internal thermal bypass valve, so do not check the level at all until TFT has reached 70 C (158 F).
• Case plug torque: 22 N.m (16 lb-ft).

An under-filled 10L90 causes premature component wear. An over-filled one pushes fluid out the vent tube, which Chevrolet Performance warns can start a fire, and causes foaming and pump cavitation on the way there.

---

The 10L90's Defining Failure: The Feed Limit Low Valve

If you remember one thing about the ten-speed, remember this valve. In March 2025 GM issued Safety Recall N242480630, filed with NHTSA as campaign 25V148000, covering 90,081 vehicles: 2019–2020 Cadillac CT6, 2020–2021 CT4 and CT5, and 2020–2022 Chevrolet Camaro. NHTSA's own language is short and ugly — the transmission "may become damaged internally and cause the wheels to lock-up." Owner letters went out April 7, 2025, and the remedy is transmission control module monitoring software, not hardware.

Sonnax traced the mechanism, and it is worth understanding because the recall software does not repair it. The feed limit low valve sits in the boost ball circuitry, but it has a second job: limiting solenoid feed to the default disable (DD) on-off solenoid. That DD solenoid controls two valves in the upper valve body, the default enable valve and the default select valve. The DD solenoid is powered on at all times and only turns off during a default event, at which point those two valves deliberately route line pressure into the exhaust backfill circuit and apply the A, B, C and D clutches through the exhaust ports — second gear, by design, as a power-down limp strategy.

Now wear the bore. Feed pressure to the DD solenoid drops, the default enable and default select valves stop being held where they belong, and the transmission starts doing that limp-mode trick when nobody asked it to. Sonnax reports two classic complaints: an upshift into seventh or eighth that suddenly drops the vehicle to second, or the drive wheels abruptly trying to lock. It is a heat-dependent failure. Let the vehicle cool and it will usually behave, which is exactly why so many of these get handed back undiagnosed.

Diagnosing it on the bench

Two things point at it hard. First, DTC P0747 — in plain terms, the E clutch is on when it should be off. Second, vacuum testing. Test the balance end of the feed limit low valve, which is the location that wears most, and compare it against the feed limit high valve; the two are the same valve train with different springs, so their readings should match. Per Sonnax, a typical healthy reading is 18 in-Hg or higher. If the bore tests below 8 in-Hg, a drop-in valve will not make up the difference and you are into oversized valve kits and bore reconditioning.

Do not let one of these keep driving while you think about it. Sonnax is explicit that extended operation in this condition will destroy clutch packs, because the wear intermittently binds the unit up. Sonnax also notes the feed limit low valve can wear out in as few as 35,000 miles on 10L1000 units doing tow and heavy-trailer work — a fair warning for any hard-worked ten-speed.

GM's hardware answer

A redesigned Generation 3 service valve body came into production June 19, 2025 under part numbers 24071206 (3.0L diesel) and 24071207 (5.3L and 6.2L gas), for ETRS models with electronic park control. It is easy to identify: the TCC solenoid moved to the case side of the valve body and the lower casting is gone, with the default disable solenoid now mounted to the lower cover plate. The feed limit high valve train has been eliminated entirely, and the default valves got a redesigned oil circuit that prevents the wheel lockup scenario even when solenoid feed to the DD solenoid goes low. It back-services the Gen 2 units despite looking nothing like them; verify fitment by VIN before ordering.

---

The 8L90's Defining Failure: Signal Accumulator Bore Wear

The eight-speed fails somewhere else entirely, and the failure is easy to misread as a solenoid problem. The 8L90 runs seven variable force solenoid circuits. Every one of them carries a small aluminum signal accumulator piston whose only job is to damp pressure and flow instability so the related control valve behaves. Those pistons are extremely active, and they wear the cast aluminum bore they ride in.

Here is why that matters more than it sounds. These are low-flow solenoids. They are not applying anything; they are supplying just enough signal pressure to stroke a clutch control valve or, in one case, the pressure regulator valve. The control valves do the heavy lifting. So any leakage past a worn signal accumulator bore has an outsized effect on the shift it feeds. Sonnax reports the S3 and S4 locations are generally the worst.

The trap is what happens next. GM does not sell the 8L90 solenoids separately, because each solenoid is individually flow-rated and that data is programmed into the TCM's control algorithm; swapping a single solenoid off a core would falsify the programming the TCM relies on. So shift and solenoid codes routinely get "fixed" with a new valve body when the actual defect is a worn aluminum bore that a piston kit resolves.

Testing it correctly

Vacuum test the signal accumulator pistons with the spring still in place, putting the test port opening in the center of the spring. Gauge pins work if you do not have a vacuum tester. One material note that gets skipped: use an aluminum replacement piston, not steel. Different thermal expansion rates open bore-to-piston clearance at operating temperature, which puts you right back where you started once the unit warms up. When the repair is done, reset the adaptive tables and run a relearn, or the customer will be back with shift complaints anyway.

---

Interchange: What Actually Swaps

Almost nothing, and the case is the reason. The 8L90 uses a two-piece main case with the bellhousing integrated into the main case — a package GM deliberately sized to occupy the same space as the outgoing 6L80 and 6L90. The 10L90 uses a case with an integral bell housing built for powertrain stiffness, with the pump hung off-axis low in the unit. They are not the same envelope, they are not the same length, and the internal hard parts do not cross.

Controls do not cross either. The 10L90's TCM is externally mounted and reads three speed sensors. Chevrolet Performance states outright that each 10L90 controller is flashed with the solenoid characteristics of the specific transmission it shipped with, that controllers and transmissions are matched, and that you cannot mix them. On the 8L90 side the same principle applies through the individual solenoid flow ratings held in the TCM. Buying a bare unit off a pallet and expecting your existing controller to run it is how these swaps die.

What does cross is diagnostic technique. Both are clutch-to-clutch units whose complaints trace back to low-flow signal circuits and worn aluminum bores, and both reward vacuum testing over parts swapping.

---

Which One Is Stronger

On the published ratings the 8L90 edges it: 665 lb-ft maximum gearbox torque and a 22,500 lb GCVW rating, against 650 lb-ft for the Chevrolet Performance 10L90. In practice both sit in the same class of holding capacity, and the sensible way to choose between them is by failure profile and by what the vehicle already has.

The 10L90 is the better transmission to drive. Wider spread, three overdrives, quicker shifts, lower cruise rpm. It also carries a documented safety recall on a valve that can wear out well inside 100,000 miles and, when it does, produces a genuinely dangerous failure. If you are buying or building one, budget for the feed limit low valve up front and put the Gen 3 valve body on the table before the customer's first complaint, not after.

The 8L90 is the simpler unit to own. One fewer clutch, fewer valves, a well-mapped wear pattern in the signal accumulator bores, and no safety recall hanging over it. It is fussy about fluid and it will punish a shop that treats a signal circuit complaint as a solenoid failure, but the repair is cheap and the diagnosis is repeatable.

---

FAQ

---

Related Guides

• 6L80 vs 8L90: Differences, Interchange, and Which Is Stronger
• 6L90 vs 8L90: Differences and Interchange
• 8L90 Pressure Testing: Ports, Specs and Procedure
• 8L90 Rebuild Wear Points
• GM 10L80 Problems and the Valve Body Recall
• 10L80 Shift Solenoid Symptoms