Why you may belief TechRadar
We spend hours testing each services or products we evaluation, so that you might be certain you’re shopping for the most effective. Find out more about how we test.
Nvidia GeForce RTX 5070: Two-minute evaluation
Numerous guarantees had been made in regards to the Nvidia GeForce RTX 5070, and in some slim sense, these guarantees are fulfilled with Nvidia’s mainstream GPU. However the gulf between what was anticipated and what the RTX 5070 really delivers is just too extensive a spot to bridge for me and the legion of players and fans on the market who will not have the ability to afford—and even discover, frankly—Nvidia’s best graphics cards from this technology.
Launching on March 5, 2025, at an MSRP of $549 / £549 / AU$1,109 within the US, UK, and Australia, respectively, this is perhaps one of many few Nvidia Blackwell GPUs you will discover at MSRP (together with accessible inventory), however just for lack of considerable demand. Because the middle-tier GPU in Nvidia’s lineup, the RTX 5070 is supposed to have broader enchantment and extra accessible pricing and specs than the enthusiast-grade Nvidia GeForce RTX 5090, Nvidia GeForce RTX 5080, and Nvidia GeForce RTX 5070 Ti, however of all of the playing cards this technology, that is the one which appears to have the least to supply potential patrons over what’s already in the marketplace at this value level.
That is to not say there may be nothing to commend this card. The RTX 5070 does rise up to native Nvidia GeForce RTX 4090 efficiency in some video games because of Nvidia Blackwell’s unique Multi-Body Era (MFG) know-how. And, to be honest, the RTX 5070 is a considerable enchancment over the Nvidia GeForce RTX 4070, so at the very least in direct gen-on-gen uplift, there’s a roughly 20-25% efficiency achieve.
However this card is a far, far cry from the promise of RTX 4090 performance that Nvidia CEO Jensen Huang presented on stage at CES 2025, even with the qualifier that such an achievement could be “unimaginable with out synthetic intelligence,” which means a heavy reliance on DLSS 4 and MFG to get this card over the road.
If we’re simply speaking framerates, then in some very slim instances this card can do this, however at 4K with ray tracing and cranked-up settings, the enter latency for the RTX 5070 with MFG might be noticeable relying in your settings, and it may possibly turn into distracting. Nvidia Reflex helps, however in the event you take RTX 4090 efficiency to imply the identical expertise because the RTX 4090, you merely will not get that with MFG, even in the 80 or so games that support it currently.
Add to all this the truth that the RTX 5070 barely outpaces the Nvidia GeForce RTX 4070 Super whenever you take MFG off the desk (which will be the case for the overwhelming majority of video games performed on this card) and you actually do not have something to indicate for the additional 30W of energy this card pulls down over the RTX 4070 Tremendous.
With the RTX 5070 coming in at lower than 4 p.c sooner in gaming with out MFG than the non-OC RTX 4070 Tremendous, and roughly 5% sooner general, that signifies that the RTX 5070 is basically a stock-overclocked RTX 4070 Tremendous, performance-wise, with the added function of MFG. An overclocked RTX 4070 Tremendous would possibly even match or exceed the RTX 5070’s general efficiency in all however a handful of video games, and that does not even contact upon AMD‘s numerous choices on this value vary, just like the AMD Radeon RX 7900 GRE or AMD’s upcoming RX 9070 XT and RX 9070 playing cards.
On condition that the RTX 4070 Tremendous remains to be typically accessible in the marketplace (at the very least in the meanwhile) at a value the place you are prone to discover it for lower than accessible RTX 5070 playing cards, and competing AMD playing cards are sometimes accessible for much less, simpler to seek out, and supply roughly the identical stage of efficiency, I actually wrestle to seek out any motive to suggest this card, even with out the questionable-at-best advertising for this card to bitter my emotions about it.
I caught loads of flack from fans for praising the RTX 5080 regardless of its 8-10% efficiency uplift over the Nvidia GeForce RTX 4080 Super, however on the stage of the RTX 5080, there isn’t a actual competitors and you are still getting the third-best graphics card in the marketplace with a noticeable efficiency enhance over the RTX 4080 Tremendous for a similar MSRP. Was it what fans needed? No, nevertheless it’s nonetheless a unbelievable card with few friends, and the bottom efficiency of the RTX 5080 was so good that the latency downside of MFG simply wasn’t a problem, making it a robust value-add for the cardboard.
You simply cannot declare that for the RTX 5070. There are just too many different choices for players to think about at this value level, and MFG simply is not a robust sufficient promoting level at this efficiency stage to maneuver the needle. If the RTX 5070 is the one card you’ve got accessible to you for buy and also you want an amazing 1440p graphics card and might’t anticipate one thing higher (and also you’re solely paying MSRP), you then’ll in the end be proud of this card. However the Nvidia GeForce RTX 5070 might have and may have been so a lot better than it in the end is.
Nvidia GeForce RTX 5070: Worth & availability
- How a lot is it? MSRP/RRP beginning at $549 / £549 / AU$1,109
- When are you able to get it? The RTX 5070 goes on sale on March 5, 2025
- The place is it accessible? The RTX 5070 will likely be accessible within the US, UK, and Australia at launch
The Nvidia GeForce RTX 5070 is offered beginning March 5, 2025, with an MSRP of $549 / £549 / AU$1,109 within the US, UK, and Australia, respectively.
This places it on the similar value as the present RTX 4070 MSRP, and barely lower than that of the RTX 4070 Tremendous. It is also the identical MSRP because the AMD’s RX 7900 GRE and upcoming RX 9070, and barely cheaper than the AMD RX 9070 XT’s MSRP.
The comparatively low MSRP for the RTX 5070 is without doubt one of the vibrant spots for this card, in addition to the existence of the RTX 5070 Founders Version card, which Nvidia will promote straight at MSRP. It will at the very least put one thing of an anchor on the cardboard’s value within the face of scalping and basic value inflation.
Nvidia GeForce RTX 5070: Specs
- GDDR7 VRAM and PCIe 5.0
- Larger energy consumption
- Nonetheless simply 12GB VRAM, and fewer compute items
Header Cell – Column 0 |
RTX 5070 |
RTX 4070 Tremendous |
RTX 4070 |
---|---|---|---|
Course of Node |
TSMC N4P |
TSMC N4 |
TSMC N4 |
Transistor Rely (Billion) |
31.1 |
35.8 |
35.8 |
Compute Items |
48 |
56 |
46 |
Shaders |
6,144 |
7,168 |
5,888 |
RT Cores |
48 |
56 |
46 |
Tensor Cores |
192 |
224 |
184 |
Render Output Items |
80 |
80 |
64 |
Cache (MB) |
48 |
48 |
36 |
Base Clock (MHz) |
2,325 |
1,980 |
1,920 |
Enhance Clock (MHz) |
2,512 |
2,475 |
2,475 |
Reminiscence Clock (MHz) |
1,750 |
1,313 |
1,313 |
Reminiscence Sort |
GDDR7 |
GDDR6X |
GDDR6X |
Reminiscence Pool (GB) |
12 |
12 |
12 |
Reminiscence Interface (bits) |
192 |
192 |
192 |
Efficient Reminiscence Velocity (Gbps) |
28 |
21 |
21 |
Reminiscence Bandwidth (GB/s) |
672.0 |
504.2 |
504.2 |
PCIe Interface |
5.0 |
4.0 |
4.0 |
TGP (W) |
250 |
220 |
200 |
Advisable PSU (W) |
600 |
550 |
550 |
Energy Connector |
1 x 16-pin |
1 x 16-pin |
1 x 16-pin |
The Nvidia GeForce RTX 5070 is a combined bag in terms of specs. On the one hand, you’ve got superior know-how like the brand new PCIe 5.0 interface and new GDDR7 VRAM, each of which seem nice on paper.
Alternatively, it appears like each different spec was configured and tweaked to guarantee that it compensated for any efficiency profit these applied sciences would impart to maintain the general package deal kind of the identical because the earlier technology GPUs.
For example, whereas the RTX 5070 sports activities sooner GDDR7 reminiscence, it would not develop the VRAM pool past 12GB, not like its rivals. If Nvidia hoped that the sooner reminiscence would make up for preserving the quantity of VRAM the identical, it solely makes a modest improve within the variety of compute items within the GPU (48 in comparison with the RTX 4070’s 46), and a noticeable lower from the RTX 4070 Tremendous’s (56).
No matter efficiency positive factors the RTX 5070 makes with its sooner reminiscence, then, is totally neutralized by the bigger variety of compute items (together with the requisite variety of CUDA cores, RT cores, and Tensor cores) within the RTX 4070 Tremendous.
The bottom clock on the RTX 5070 is notably increased, however its enhance clock is barely barely elevated, which is in the end the place it counts whereas taking part in video games or operating intensive workloads.
Likewise, no matter positive factors the extra superior TSMC N4P node affords the RTX 5070’s GPU over the TSMC N4 node of its predecessors appears to be eaten up by the chopping down of the die. If there was an influence or price motive for this, I do not know, however I feel that this determination is what in the end sinks the RTX 5070.
It looks as if each determination was made to maintain issues proper the place they’re moderately than transfer issues ahead. That may be acceptable, truthfully, if there was another main profit like a drastically diminished energy draw or a lot cheaper price (I’ve argued for both rather than pushing for more performance every gen), however by some means the RTX 5070 manages to tug down an additional 30W of energy over the RTX 4070 Tremendous and a full 50W over the RTX 4070, and the value is barely barely decrease than the RTX 4070 was at launch.
Lastly, it is a PCIe 5.0 x16 GPU, which signifies that if in case you have a motherboard with 16 PCIe lanes or much less, and also you’re utilizing a PCIe 5.0 SSD, considered one of these two parts goes to get nerfed right down to PCIe 4.0, and most motherboards default to prioritizing the GPU.
You would possibly have the ability to set your PCIe 5.0 precedence to your SSD in your motherboard’s BIOS settings and put the RTX 5070 into PCIe 4.0, however I have never examined how this may have an effect on the efficiency of the RTX 5070, so be aware that this is perhaps a problem with this card.
Nvidia GeForce RTX 5070: Design
- No dual-pass-through cooling
- FE card is identical dimension because the RTX 4070 and RTX 4070 Tremendous FE playing cards
The Nvidia GeForce RTX 5070 Founders Version seems to be equivalent to the RTX 5090 and RTX 5080 that preceeded it, however with some very key variations, each inside and outside.
Among the best issues in regards to the RTX 5090 and RTX 5080 FE playing cards was the progressive twin pass-through cooling resolution on these playing cards, which improved thermals a lot that Nvidia was in a position to shrink the scale of these playing cards from the gargantuan bricks of the final technology to one thing way more manageable and sensible.
It could have been good to see what such an answer might have executed for the RTX 5070, however possibly it simply wasn’t attainable to engineer it so it made any sense. Regardless, it is unlucky that it wasn’t an possibility right here, despite the fact that the RTX 5070 is hardly unwieldy (at the very least for the Founders Version card).
In any other case, it sports activities the identical 16-pin energy connector placement because the RTX 5090 and RTX 5080, so 90-degree energy connectors will not match the Founders Version, although you’ll have higher luck with most, if not all, AIB accomplice playing cards which is able to probably follow the identical energy connector placement of the RTX 40 collection.
The RTX 5070 FE will simply match inside even a SFF case with ease, and its lighter energy draw signifies that even when it’s a must to depend on the included two-to-one cable adapter to plug in two free 8-pin cables out of your energy provide, it would nonetheless be a reasonably manageable affair.
Lastly, like all of the Founders Version playing cards earlier than it, the RTX 5070 has no RGB, with solely the white backlight GeForce RTX brand on the highest fringe of the cardboard to offer any ‘aptitude’ of that kind.
Nvidia GeForce RTX 5070: Efficiency
- Virtually no distinction in efficiency over the RTX 4070 Tremendous with out MFG
- Utilizing MFG can get you native RTX 4090 framerates in some video games
- Considerably sooner efficiency over the RTX 4070
A notice on my information
The charts proven beneath supply the newest information I’ve for the playing cards examined for this evaluation. They might change over time as extra card outcomes are added and playing cards are retested. The ‘common of all playing cards examined’ consists of playing cards not proven in these charts for readability functions.
Boy howdy, right here we go.
The very best factor I can say in regards to the efficiency of this card is that it’s simply barely the best 1440p graphics card in the marketplace as of this evaluation, and that DLSS 4’s Multi Body Era can ship the sort of framerates Nvidia guarantees in these video games the place the know-how is offered, both natively or by the Nvidia App’s DLSS override function.
Each of these statements include loads of caveats, although, and the RTX 5070 would not make sufficient progress from the final gen to make a compelling case for itself performance-wise, particularly since its signature function is barely accessible in a smattering of video games in the meanwhile.
On the artificial facet of issues, the RTX 5070 seems to be sturdy in opposition to the cardboard it is changing, the RTX 4070, and usually affords about 25% higher efficiency on artificial benchmarks like 3DMark Metal Nomad or Velocity Manner. It additionally has increased compute efficiency in Geekbench 6 than its direct predecessor, although not be as drastic a margin (about 10% higher).
In comparison with the RTX 4070 Tremendous, nevertheless, the RTX 5070’s efficiency is barely about 6% higher general, and solely about 12% higher than the AMD RX 7900 GRE’s general artificial efficiency.
Once more, a win is a win, nevertheless it’s a lot nearer than it ought to be gen-on-gen.
The RTX 5070 runs into comparable points on the artistic facet, the place it solely outperforms the RTX 4070 Tremendous by about 3% general, with its greatest efficiency coming in PugetBench for Creators’ Adobe Premiere benchmark (~13% higher than the RTX 4070 Tremendous), however faltering considerably with Blender Benchmark 4.3.0.
This is not too stunning, because the RTX 5070 hasn’t been launched but and GPUs are inclined to carry out higher in Blender a number of weeks or months after the cardboard’s launch when the devs can higher optimize issues for brand new releases.
All in all, for this class of playing cards, the RTX 5070 is a stable alternative for many who would possibly wish to dabble in artistic work with out a lot of a monetary dedication, however actual professionals are higher off with the Nvidia GeForce RTX 5070 Ti in the event you’re seeking to improve with out spending a fortune.
It is with gaming, although, the place the actual heartbreak comes with this card.
Technically, with simply 12GB VRAM, this is not a 4K graphics card, however each the RTX 4070 Tremendous and RTX 5070 are sturdy sufficient playing cards that you could get playable native 4K in just about each sport as long as you by no means, ever contact ray tracing, international illumination, or the like. Sadly, each playing cards carry out roughly the identical beneath these circumstances at 4K, with the RTX 5070 pulling right into a slight >5 fps lead in just a few video games like Returnal and Dying Gentle 2.
Nevertheless, in some titles like F1 2024, the RTX 4070 Tremendous really outperforms the RTX 5070 when ray tracing is turned on, or when DLSS is ready to balanced and with none Body Era. General and throughout totally different setting configurations, the RTX 5070 solely musters a roughly 4.5% higher common FPS at 4K than the RTX 4070 Tremendous.
It is just about the identical story at 1440p, as nicely, with the RTX 5070 outperforming the RTX 4070 Tremendous by about 2.7% throughout configurations at 1440p. We’re actually within the realm of what a great overclock can get you on an RTX 4070 Tremendous moderately than a generational leap, regardless of all of the next-gen specs that the RTX 5070 brings to bear.
OK, however what in regards to the RTX 4090? Can the RTX 5070 with DLSS 4 Multi Body Era match the native 4K efficiency of the RTX 4090?
Sure, it may possibly, at the very least in the event you’re solely involved with common FPS. The one sport with an in-game benchmark that I can use to measure the RTX 5070’s MFG efficiency is Cyberpunk 2077, and I’ve included these outcomes right here, however in Indiana Jones and the Nice Circle and Dragon Age: Veilguard (utilizing the Nvidia App’s override operate) I just about discovered MFG to carry out persistently as promised, delivering considerably sooner FPS than DLSS 4 alone and touchdown within the ballpark of the place the RTX 4090’s native 4K efficiency finally ends up.
And as long as you keep distant from ray tracing, the bottom framerate at 4K will likely be excessive sufficient on the RTX 5070 that you just will not discover an excessive amount of, if any, latency in lots of video games. However whenever you flip ray tracing on, even the RTX 5090’s native body price tanks, and it is these baseline rendered frames that deal with modifications based mostly in your enter, and the three AI-generated frames based mostly on that preliminary rendered body do not consider no matter enter modifications you’ve got made in any respect.
As such, despite the fact that you may rise up to 129 FPS at 4K with Psycho RT and Extremely preset in Cyberpunk 2077 on the RTX 5070 (blowing well past the RTX 5090’s native 51 common FPS on the Extremely preset with Psycho RT), solely 44 of the RTX 5070’s 129 frames per second are reflecting lively enter. This results in a state of affairs the place your sport seems to be like its flying by at 129 FPS, however feels prefer it’s nonetheless a sluggish 44 FPS.
For many video games, this is not going to be a deal breaker. Whereas I have never tried the RTX 5070 with 4x MFG on Passable, I am completely constructive I cannot really feel the distinction, as it is not the sort of sport the place you want quick reflexes (aside from coping with the effing Stingers), however Marvel Rivals? You are going to really feel it.
Nvidia Reflex undoubtedly helps take the sting off MFG’s latency, nevertheless it would not utterly remove it, and for some video games (and players) that’s going to matter, leaving the RTX 5070’s MFG expertise an excessive amount of of a combined bag to be a categorical promoting level. I feel the hate directed at ‘pretend frames’ is wildly overblown, however within the case of the RTX 5070, it is not solely with out advantage.
So the place does that depart the RTX 5070? General, it is the most effective 1440p card in the marketplace proper now, and it is comparatively low MSRP makes it the most effective worth proposition in its class. It is also more likely that you will really have the ability to discover this card at MSRP, making the query of worth extra than simply educational.
For many players on the market, Multi Body Era goes to be nice, and as long as you go straightforward on the ray tracing, you will in all probability by no means run into any sensible latency in your video games, so in these situations, the RTX 5070 would possibly really feel like black magic in a circuit board.
However my downside with the RTX 5070 is that it’s completely not the RTX 4090, and for the overwhelming majority of the video games you are going to be taking part in, it by no means will likely be, and that is basically what was promised when the RTX 5070 was introduced. As an alternative, the RTX 5070 is an RTX 4070 Tremendous with just a few video games operating MFG slapped to its facet that appear to be they’re taking part in on an RTX 4090, however might or might not really feel like they’re, and that is simply not adequate.
It is not what we had been promised, not by an extended shot.
Must you purchase the Nvidia GeForce RTX 5070?
Class |
Notes |
Rating |
---|---|---|
Worth |
The most cost effective of the Blackwell GPUs, you would possibly really have the ability to discover this one at MSRP. |
4 / 5 |
Specs |
GDDR7 and PCIe 5.0 look nice on paper, however the whole card appears designed to ensure these two options are a whole non-factor. Additionally, please cease making 12GB RTX 4070s. Reminiscence is not that costly. |
2.5 / 5 |
Design |
There isn’t any twin pass-through cooling on this card and it is just about inoffensive on each entrance, however would not break the mildew in any manner both. |
3.5 / 5 |
Efficiency |
With out DLSS 4 Multi Body Era, this card affords subsequent to no actual efficiency positive factors over the RTX 4070 Tremendous, and only a few video games have MFG. Those who do, it really works nice, nevertheless it’s not almost sufficient. |
2.5 / 5 |
Closing rating |
The RTX 5070 is a really laborious card to suggest, particularly when there are such a lot of competing playing cards at this value that supply comparable efficiency. |
3.13 / 5 |
Purchase the Nvidia GeForce RTX 5070 if…
Do not buy it if…
Additionally contemplate
How I examined the Nvidia GeForce RTX 5070
- I spent a few week with the RTX 5070
- I used my full GPU testing suite to research the cardboard’s efficiency
- I examined the cardboard in on a regular basis, gaming, artistic, and AI workload utilization
Take a look at System Specs
Listed below are the specs on the system I used for testing:
Motherboard: ASRock Z790i Lightning WiFi
CPU: Intel Core i9-14900K
CPU Cooler: Gigabyte Auros Waterforce II 360 ICE
RAM: Corsair Dominator DDR5-6600 (2 x 16GB)
SSD: Essential T705
PSU: Thermaltake Toughpower PF3 1050W Platinum
Case: Praxis Wetbench
I spent a few week testing the Nvidia GeForce RTX 5070, utilizing it as my essential workstation GPU for artistic content material work, gaming, and different testing.
I used my up to date testing suite together with trade commonplace instruments like 3DMark and PugetBench for Creators, in addition to built-in sport benchmarks like Cyberpunk 2077, Civilization VII, and others.
I’ve reviewed greater than 30 graphics playing cards for TechRadar within the final two and a half years, in addition to extensively testing and retesting graphics playing cards all year long for options, evaluation, and different content material, so you may belief that my evaluations are based mostly on expertise and information, in addition to my need to ensure you get the most effective GPU to your hard-earned cash.
- Initially reviewed March 2025