ALL ATX v3.x PSUs Compared
Best Three PSUs
| Section | #1 | #2 | #3 |
| Best Overall Performance (115V) | NZXT C1500 | Seasonic Prime Noctua Edition TX-1600 | Seasonic Prime PX-1600 |
| Best Overall Performance (230V) | NZXT C1500 | Seasonic Prime Noctua Edition TX-1600 | Seasonic Prime TX-1600 |
| Highest Average Efficiency (115V) | Seasonic Prime TX-1300 | FSP Hydro Ti Pro 1000W | Seasonic Prime Noctua Edition TX-1600 |
| Highest Average Efficiency (230V) | Seasonic Prime TX-1300 | Seasonic Prime Noctua Edition TX-1600 | FSP Hydro Ti Pro 1000W |
| Highest Average Efficiency – 20-80W (115V) | Cooler Master GX II 850 Gold | Cooler Master GX II 750 Gold | Cooler Master GX III 750 Gold |
| Highest Average Efficiency – 20-80W (230V) | Cooler Master GX II 850 Gold | Cooler Master GX II 750 Gold | Cooler Master GX III 650 Gold |
| Tightest Load Regulation 12V (115V) | Super Flower Leadex VII PRO 850W | Corsair SF850L | Super Flower Leadex VII PRO 1000W |
| Lowest Ripple 12V (115V) | Asus Rog-Strix-750G-Aura-Gaming | Asus Rog-Strix-1000G-Aura-Gaming | Asus Rog-Strix-1200G-Aura-Gaming |
| Longest Hold-up Time (230V) | Cougar Polar X2 1200W | Cooler Master V750i | NZXT C1000 Gold |
| Lowest Average Noise (115/230V) | Cooler Master X Silent Max Platinum 1300 | Sharkoon Rebel P20 750 | Seasonic Prime TX-1300 |
| Best ATX v3.x Transient Response (200% Load) | Segotep GM750W | Lian Li EDGE 850 | Corsair SF850 |
The Salcomp ATX 1600W has the best overall performance, but since this unit is unavailable on the market, I don’t include it in the table above. The same goes for the Zhong Yuan Power GX850, which has the longest hold-up time.
Overall Performance 115/230V
 |
 |
Average Efficiency 115/230V
 |
 |
Average Efficiency 20-80W Range – 115/230V
 |
 |
Average Efficiency Ultra Light Loads (2%) – 115V
Hold-Up Time – 230V
Load Regulation & Ripple 12V
 |
 |
Average Noise 115/230V
 |
 |
ATX v3.x Transient Response Testing – 200% Load
 |
 |
If you want to support my work, please consider purchasing through the following affiliate links. There won’t be an extra charge, and Amazon will pay me a small commission from every sale, helping me to continue doing what I do.
Pages:
Is it possible to test the cooler master mwe v3. We have it for 65 euros 650w. I hope it’s a great buy like the v2
if I can find a recent sample why not.
Hello! I recently bought an NZXT C850 ATX 3.1 (2024). I saw the higher wattage versions were reviewed but is this one also still very good?
It was on sale for $110 USD, it looked very promising. Corsair RMx is not on sale rn and the Shift version isn’t ideal for my case plus still a bit more expensive despite also being on sale.
It is good yes.
Asrock t-1300 taichi review soon?
1650 first 🙂
C1500 was #1 in both best overall categories if i recall correctly 🙁
yeap we added some more factors in our overall performance algorithm so there are some changes. We contantly adapt to the times and improve our performance algo.
Hello Aris.
Have you tested the 750w version of the Loki? I could not find it on your site or youtube channel. Has it been considered for this list?
Thank you.
Check the Cybenetics Database
Should I get the:
Corsair SF850L ($145) or
ROG LOKI SFX-L 750W ($165)?
Thanks!
The SF850L
Thanks!
Should I get the Corsair SF850L (145$) or the Corsair SF1000 (215$) ? Thanks!
Hey there!
I can’t decide between XPG Core Reactor II and Core Reactor II VE (both of them at 750W). They can be find on various lists here (650-750-850W etc), so it is decent, however I heard that VE is a little cheaper, so that means the baseline model would be a better choice, per se?
Thank you in advance!
If the price difference is low, get the Reactor II
Thank you so much, last question if you don’t mind: would you recommend this or Be quiet! PURE POWER 12 M 750W? (same price too)
Good Day. I just wanted to know if you have tested the Super Flower Leadex III GE 80 Plus Gold models. 1000w or 1300w. I would like to know your thoughts on it.
Why is corsair ax1600i not listed here? Would you still recommend it over the listed options?
it isn’t ATX v3.x compliant. Still, a great PSU.
Thank you, i noticed its not ATX 3.x after i wrote it. But what confuses me is, what is an actual minimum ATX requirement for modern high performance gaming pcs. As you pointed out, ATX 3.1 is actually worse than ATX 3.0 as it worsens the hold up time requirement. Is there something like a minimum ATX version requirement for PSUs that you recommend.
Sorry, sir, I’m a bit confused with your list. It mentions Super Flower Leadex VII XP 1000W and Super Flower Leadex VII Pro 1000W. On the website, I only found the Pro platinum version with model (SF-1000F14XP). Could you tell me the model number for the XP version you tested? The quality difference between the two PSUs is significant. I’m interested in the XP version, but I couldn’t find it anywhere.
I also couldn’t find your review results in the PSU 1000W list for the XP model.
Exactly. I think there are many people who want to know.
Super Flower kind of messed up its model numbers. I asked them to reply to this.
Great work as always. I have been really happy with my Leadex XG 850w you recommended. With Black Friday I ended up ordering the Corsair Rm850X 3.1 for a second build since it was actually cheaper and on sale, 154.99 vs 124.99, though the Leadex includes 3 free case fans. 750x wasn’t on sale so it was $6 more to go for the 850x over the 750x.
Amazon now has a link if you want to replace the RM750e which looks to be out of stock now, https://www.amazon.com/CORSAIR-RM750x-Modular-Low-Noise-Supply/dp/B0DJ1M18CY for the 750 and https://www.amazon.com/CORSAIR-RM850x-Modular-Low-Noise-Supply/dp/B0DJ1JL3MK for the 850. As of right now both of these specify the ATX 3.1 version and the 3.0 version has a different link.
Hey!
First of all, thank you for the amazing work that you do.
Right now I am looking for a new PSU and have narrowed it down to the RM1000x (2024) and the NZXT C1000 (atx 3.1).
They are both at the same price right now (150€) so its pretty hard to choose. Which one would you go for? I like that they are both low noise. The C1000 has better light load efficiency, but the rm1000x seems like the better allrounder.
I have the exact same question! I find it hard to choose between those two.
I’m also hoping for an answer to this question!
RM1000x
Thank you! 🙂
Sorry, im new here and abit confuse.
The chart clearly said nzxt is in the top 3 best, but why corsair is the recommended one?
Thanks
Sorry im new here and abit confuse,
The chart clearly said that nzxt are one of the best top 3 but why corsair are the one recommended?
Thankyou
They asked me what I would buy, not what the charts say. There is a big difference. If you can read the charts, why still ask what to buy?
Thanks a lot!
Hi! I searched on your website for a review of NZXT C850 ATX3.1 but could not find it, instead I found those of the C1000, C1200 and C1500, so I was wondering where C850 data used in the article Best 850W PSU Picks came from? Maybe Cybenetics?
Can I refer to C1000 review for C850 as well (excluding max power of course)?
Besides power, are there any other differences between C850 and C1000?
I’m currently looking for a new PSU, but I’m undecided between these 3 models:
NZXT C850 ATX3.1
Corsair RM850X ATX3.1
LC-Power LC850P ATX3.0
Now in my country they all have the same price about 130€, from the article it seems C850 is a better choice over the RM850X is that so?
It looks LC-Power products are not available in your country, the model quoted above is a 80+Platinum unit and is based on a Super Flower platform (possibly Leadex IV) with Teapo capacitors.
Do you have any advice for me?
Thank you for your amazing work
It’s incredible the work they do, all the information they bring to the community is invaluable. I have a question about whether to buy a be quiet! Pure Power 12, 850 W ATX 3.1 PCI 5.0 or an MSI A850GL ATX 3.0. The MSI is about 14 dollars cheaper than the be quiet
@crmaris: Great work on all of these reviews. I spent several hours reading on here yesterday. 🙂
I’m considering a PSU upgrade while I’m upgrading my other components. I do prefer low noise, so the FSP Hydro Ti Pro 1000W looks like a pretty great choice. The Corsair RM1000x ATX 3.1 is a lot cheaper currently though due to Black Friday related sales currently going on.
Corsair RM1000x ATX 3.1:
$168 (on sale) incl. 25% VAT ($134.4 without VAT), roughly $58 off normal local price
FSP Hydro Ti Pro 1000W:
$278 incl. 25% VAT ($222.4 without VAT)
Would you say that going for the FSP unit is worth it still? Is the noise difference noticeable in practice? It looks like it could be. On the other hand, I’m currently using a Corsair RM750x from 2016 so I’m not really sure how that would stack up noise wise.
Alternatively, I could also just continue to use my old Corsair RM750x from 2016 for a while longer while I wait for newer GPUs to come out early next year and hope that the price of the FSP unit has dropped a bit. It’s not available in particularly many stores targeting the Swedish market though, only 3 stores from Denmark (for some reason).
Any thoughts?
If you want the best the Hydro Ti Pro is at another category. That said, the RM1000x is also very good.
Right. Getting the best is of course a nice thought, and at normal pricing it would be a no brainer since the price difference wouldn’t be nearly as large as it is now. I’m just trying to determine whether the difference is immediately noticable or not in practice?
Do you happen to know how the older Corsair RMx models would compare noise wise by any chance? I saw your review of the older RM850x model where you complained about fan noise. It seems that models before that one also used FDB fans so they probably didn’t have the fan noise issue you mentioned.
Based on a copy of the user manual for my RM750x 2015 model, I decided to do an unscientific visual comparison of the noise graphs Corsair provide. I basically scaled them to match and then overlayed the graphs.
RM1000x (2024) vs. RM750x (2015):
https://imgur.com/a/vtXxyIR
RM1000x (2024) vs. RM1000x (2015):
https://imgur.com/a/pHWkK4h
Based on this, it seems like the 2024 generation of the RM1000x is more quiet than the 2015 RM1000x until it reaches roughly 750W load, where it then overtakes the 2015 model and ends up roughly 10 dB louder at 1000W load.
When comparing my RM750x to the 2024 RM1000x, the latter is quieter for the entire output range of my 750W model.
Note that I’m assuming that they use the same measurement methodology, which quite possibly isn’t the case.
@crmaris Something I find kind of strange are the spots of higher noise at certain minor rail loads found in both your FSP graph and Cybenetics’ FSP graph (both 119V and 230V). How are we supposed to interpret that in practice?
https://www.cybenetics.com/evaluations/psus/2101/
It seems as though there are instances where certain load combinations on the minor rails cause the FSP unit to go from less than 6 dB to 30-35 dBA noise based on the 230V graph. At other combinations of load on the minor rails it’s at <6 dBA at a higher load on both the 12V and minor rails. To me, that indicates that there might be a particular minor rail that causes the fan to ramp up when loaded. Do you know what particular rail that is?
It seems like pretty weird behavior to me without knowing which rail it was, but you don't comment on it in the review so I have no way of knowing how likely I am to run into that edge case. You do however provide insight on the behavior of the CWT platform used by the RM1000x.
The minor rails in most PSUs have next to zero passive cooling, so when you apply high loads to them, the fun begins 🙂 Meaning that the fan’s speed goes high to cool them down. Once it does so, it slows down (or stops). We apply equal Amps to each of the minor rails during these tests.
I see. Thanks for the additional context and the continued responses!
The Corsair unit doesn’t show such spots because the fan is already running at that point in most cases, I presume?
Reading the Anandtech review, I noticed that the difference in noise measurements are huge. Sure, their noise floor is apparently 30.4 dbA, but they still measure it at 32 dbA at 500W and close to 44 dbA at 1000W while your testing as well as the Cybenetics report has it in the 20-30 dBA range depending on input voltage (I’m guessing in the middle somewhere as it makes little sense to me for it to be much louder when running more efficient at 230V). At 800W load, you and Cybenetics have it at less <6 dbA while they have it at roughly 38.5 dbA. Sure, your graph shows that it depends on the load on the minor rails, but their graph shows an almost linear increase of dbA values (which themselves are a logarithmic) from 500W up to 1000W.
https://www.anandtech.com/show/21148/the-fsp-hydro-ti-pro-1000w-psu-review/2
Basically, I have a hard time reconciling the two. What am I missing? 🙂
Looking at their Corsair RM750e review, they measure over 50 dbA at 750W load while Corsair themselves state roughly 26 dbA. That's a mind blowing difference to me, no matter how cherry picked you think Corsair's own results are.
They do state that they actually test the fan separately without running a load test, but I don't see how that explains the difference unless they also measure very close to the fan or something.
"In other words, instead of trying to do the impossible and measure the noise of a power supply while we are testing it, there is nothing keeping us from using a non-intrusive laser tachometer to record the speed of the fan instead. Then, we can simply test the unit on its own, with the fan hotwired to a small fanless, adjustable DC PSU that we fabricated, taking noise readings with our Extech HD600 for the RPM range of the fan and cross-referencing the two tables."
Does the guy on anandtech says how he measures noise? Because there is huge difference if you measure from 1 meter or from 20/50 cm. We follow the equivalent ISO for noise measurements in IT equipment and our noise measurement equipement costs around 25-30K euros, excluding of course the cost of the hemia-anechoic chamber. Also with 30 dBA noise floor, he can only take noise measurements at 33 dBA and above. Just saying 🙂
No, I don’t see any information about how close they measure it. It makes sense that it makes a massive difference how far away you measure it. I suppose that must be the reason. 🙂
Yeah, a noise floor that high isn’t nearly as useful for measuring the noise level of more silence focused components like these, but I also understand that anechoic chambers are a huge investment so they have to make due with what they have. Still, I wish that they would’ve provided more details about that aspect of their methodology. 🙂
The link to their methodology is the one below, so I would assume that is still how they do it:
https://www.anandtech.com/show/7820/how-we-test-psus-2014
Regarding those minor rail load situations, how likely do you think one would be to run into those often enough to be bothered by them? After all, a fan that routinely spins up and down could in some cases be more irritating than a more consistent noise (depending on frequency etc.). Hopefully that’s my last question. 😀 I’ve already taken up too much of your time. 🙂
I just realized that the Corsair RM1000x in your review almost never stopped the fan from running, even at very low 12V rail loads, or basically any load for that matter. Why is that? 🙂
Sure, if the 3.3V and 5V rails saw a lot of load it would make sense, but otherwise the fan is supposed to stay off until 500W load. That doesn’t seem to be what your unit did at all.
@blunden, sorry to hijack your thread.
Is the FSP you listed a ATX 3.1 version ? Where do you find it on sale (in USA) ? I can’t seem find any place selling the ATX 3.1 model
I don’t know whether it’s the ATX 3.1 version or not. It simply doesn’t say, which seems to match what others have experienced.
I’m not based in the US so I don’t know where it’s sold there. The Amazon referral link from the review seems to point to the updated model though.
https://amzn.to/3Wacnvk
Thanks for the reply @blunden.
The one on Amazon appears to be ATX 3.0 only.
FSP website lists the ATX 3.1 version, but doesn’t look it is being sold anywhere in US yet.
https://www.fsplifestyle.com/en/product/HydroTIPRO1000W_ATX31.html
No, it clearly says “12V-(2X6)” in the title. That’s pratically all they had to change to adhere to the ATX 3.1 spec, which is otherwise less strict than ATX 3.0.
Note that the model numbers of both listings on FSP’s site are the same. They basically started soldering on a different connector on the PSU, change the printing on the unit to reflect that as well as replace the cable. Based on what others have said, it’s a silent revision change and not actually a new model.
Judging by the July 14 2024 comment from @Mar on the actual HWBusters review, the boxes were still saying ATX 3.0 on them at that point in time despite this minor update. I wouldn’t be surprised if that was still the case.
I’d say it’s a pretty safe bet that the Amazon listing I linked earlier is the updated model. If not, you’ll simply have to return it.
I just saw TweakTown’s review of this PSU and their unit came with updated box art. It still said ATX 3.0 but it clearly mentioned 12V-2×6 on the box in multiple locations so you should in that case be able to tell if it’s the updated unit without opening the box. I imagine that would help with returns.
Hey, may I please ask, why is there no actual review of the Leadex VII 850 XG?
Thank you.
Indeed. Will see what I can do about that.
You mentioned that you added Deepcool PL750D to the best ATX 3.0 PSUs list. But I cant find it anywhere. So is the PL750D worth buying? Will it be reliable?
Hey, just wondering if the Leadex titanium 1600w ATX 3.1 would perform the same as the Leadex titanium 1600w ?
trying to decide for a threadripper 7960x and dual 3090s, Not sure if I should go for the 3.1 version since there are not much reviews yet on the 3.1 variant. Thanks!
Hello i’ve been planning to grab the Super Flower LEADEX VII Platinum PRO 1200W ATX 3.1 for my next build but i can’t find it anywhere in the US.
Coole Master MWE Gold 650 V2 Full Modular it is good PSU?
Great article. I believe the original season tx 1600 wasn’t fully atx 3.0 compliant and had some like of limitation to 1000w. Does the noctua edition have this issue?
Hi, my vertex gx-1200w jjust died (luckily without damaging my system) and i have to replace it.
Should you go for another seasonic 1200w or maybe for a super flower leadex vii pro 1200w?
Or any other suggestion will be pleased 🙂
I use Vertex GX-1200 on all of my test systems without any issues. You should give Seasonic one more chance I think 🙂
Yes, i was more for seasonic then super flower, i will follow your advice. One unfortunately device can happen to everyone.
Thanks a lot for your answer
Hello sir, which 750w would you recommend to buy in Amazon? Im trying not to get too fancy or expensive since I’ll be running it with a R5 5600 and a rx 7800 xt. Thanks in advance.
I saw NZXT C850 in some rankings. Is that 2022 ?
yes
Hi Aris,
I saw the report on Cybenetics Super Flower III 1300W (ATXv3.1) where you mentioned the unit Broke in 5VSB OCP testing. Is this something to be concerned about if I were to buy this unit for ATX 3.1?
Thank you.
It doesn’t concern you if you don’t push 5VSB too high. That said, from now on, we decided we won’t provide an ATX v3.1 pass to any unit that breaks during the evaluation of protection features.
First of all, thank you very much for this detailed review.
Can you give me a hint what would be the best pick for a Ryzen 78003DX and a 7900XTX?
So far I’m wavering between the Super Flower Leadex VII XG 850W, Super Flower Leadex VII PRO 850W or the Asus Rog-Strix-850G-Aura-Gaming.
But I’ve also seen that you often recommend the Corsair RM850x.
Thanks for your advice!
I’m so glad that I found this website a few weeks back as I was researching for my first computer build.
Thank you thank you thank you for putting the time and effort into something like this. It’s by far the best resource I’ve seen for PSU’s on the internet.
One suggestion: For the list of best power supplies, maybe you can include a way to filter them by price, or add a ranking by price too? Obviously, it makes sense to put them into tables by wattage like you already have, but I think you would gain even more traffic to this article if people could use it to help find the answer for “best 850w PSU under 100, under 150, under 200”, etc (just example price ranges)
Lastly, one question: I’m looking for an atx 3.1 850W power supply between 100-150 dollars. I’ve mostly narrowed it down to the new RM850x or the NZXT C850. Which one is a better option? Or should I consider something else?
Thank you so much!
We will try to work on something for the best PSUs; thank you for the suggestion! However, prices are not that easy to follow since they constantly change.
Both are great, personally I would go for the RM850x
Thank you, very much appreciate the reply!
I agree tracking prices can be tough, maybe to start it could be based on MSRP..
I was leaning towards the RM850x, thank you for confirming!
Hi, I’m thinking of getting Lian Li Edge 1000w for my r9 9900x and 4070 ti super and haf 700 case. Im not sure if the asus strix wont be a better choice but I do like usb hub and cables included with Lian Li. I do understand that it is included in the best table but do you recommend Edge 1000w?
I have data for it from Cybenetics, but I haven’t reviewed it and won’t so cannot advice about it.
Hi, I have a Segotep Gw750, and a Sama black power 850w . Should I keep the Sama and return the Segotep. I can’t find much on Sama.
I wouldn’t trust the SAMA PSU.
I’m looking at buying the FSP hydro ti 1000w 3.0
Now that they have updated it to have a 12v-2×6 connector on the psu side, does this make it a ATX 3.1 psu?
Does it have the sense0 and sense1 pins for open to open 0 watts?
Hello, which one would you prefer between Corsair RM850x ATX v3.1 and Asus Rog-Strix-850G-Aura-Gaming? Thank you for all the amazing work you’ve done.
Corsair RM850x
Thank you!
Hi Aris,
I know the Corsair RM850x (2021) is not on the charts as it’s not really an ATX 3.x unit but where would it roughly go with the Corsair 12vhwpr adapter? Based on the Pure Power 12M review would it go above it maybe around the Vertex/Shift %?
Thanks!
Where can I read Lian Li Edge review?
I didn’t review it, sorry.
Hello, I apologize for asking so many questions lately.When I found the hx1500i for a very reasonable price, I bought it thinking it was a good deal, but I’m confused about one thing.You said that this power supply does not meet the requirements of ATX 3.0. Should I return it? (I paid $220, everything included)
No worries. Check with Corsair. I think they must have a 12VHPWR (or 12V-2×6) cable for it, taking power from 2x 8pins. If you get that, no need to worry. No GPU so far has high enough spikes to push a PSU like that to its limits.
Of course but I don’t think the lack of ATX 3.0 is a reason to return it. Yes, I also bought an original Corsair cable Do you think I should return it?
If it covers your needs, why return it? You bought it at a good price.
Yes but I wanted real 3.1 support. But I guess I won’t find it for a long time. Thanks for all your answers and help.What motivated me to return it was reading on reddit that multiple people were having issues with the 4090 and hx1500i crashing and freezing.
Thank you for this detailed review. Asus Rog-Strix-750G-Aura-Gaming is an atx 3.0 or 3.1 I still can’t understand sorry. I know that atx 3.0 standards are harder but what i want to know is whether asus rog aura uses 2×6 or 12vhpwr? thank you in advance for the answer.
the only difference between ATX v3.0 and v3.1 is on the PSU’s connector on its modular panel, and the hold-up time. If a PSU has 12VHPWR an meets all other requirements, it is ATX v3.0. If the SAME PSU has 12V-2×6 connector, it is also ATX v3.1 compliant.
Thanks for the answer. I know that but I am looking for a real 3.1 psu for the safety factor and corsair rmx atx 3.1 is not yet available in my country. So Even though the price of this product is affordable in my country, I pass on it.
Although it was stated as 12VHPWR in the review, I was curious if they were using the current pin since Intel had issued a 3.1 certification.
Hi!
How come the “be quiet!” Pure Performance Power 12 M 1200w is listed in the bottom of the 1001-1250w page, but isn’t actually involved in any of the charts? I just bought this PSU but was hoping to see how it performed here…
I will check now that I will renew the charts.
You can also check here: https://www.cybenetics.com/index.php?option=psu-performance-database
Hey Aris.
The MSI MEG is duplicated in multiple graphs with different results. Which one should we considerate?
Thank you for your amazing work
Hi! We tested both samples and by mistake I included both. Will fix it when possible. Thank you for letting me know!