If it wasn’t a $1,000+ graphics card, I might really feel sorry for the RTX 5080. It is actually half the GPU of the large boi RTX 5090—Nvidia’s flagship GeForce Blackwell card—and but suffers from severely restricted inventory and punitively excessive pricing. Nevertheless it’s the second-tier RTX 50-series card, and might not less than declare to be the second quickest gaming GPU round, and that is received to rely for one thing, proper?
And but right here comes the RTX 5070 Ti—its supposedly cheaper, youthful sibling—kicking its shins and stealing its lunch. As a client, I am completely right here for this: a brand new GPU punching above its weight class. However it’s a little unusual contemplating Nvidia has beforehand been so tight on segmenting its playing cards purely on efficiency. I’ve additionally received a sense there’s one thing uncomfortably funky occurring with the pricing post-launch, which is bound to tarnish the RTX 5070 Ti.
Nonetheless, on the one hand, kudos to Nvidia for following up the stupidly named RTX 4070 Ti Tremendous with a comparatively specced RTX Blackwell model, and giving it a nominal $50 worth lower into the discount. However then again I am struggling to see why the RTX 5080 was hoofed out the door alongside the RTX 5090 after which adopted up with a card that is not far off it on paper, after which shreds that efficiency delta with an enormous quantity of overclocking headroom.
Yeah, the RTX 5070 Ti is that card. The overclocking hero that may nearly match the gaming efficiency of the GPU above it within the stack in the event you’re keen to dig into the headroom Nvidia has one way or the other left you to play with. And the very best bit is it is not even that arduous to do; you do not have to be some OC genius to hit 3.2 GHz+ with this GPU.
Nvidia RTX 5070 Ti: The decision
Purchase if…
✅ You could find it at MSRP: That is going to be the standing recommendation for any RTX 50-series buy for the foreseeable. Whereas inventory ranges stay expectedly tight retailers and scalpers are going to be in full impact making an attempt to make you pay nearer to $1,000 for this card.
✅ You are comfy overclocking: There’s a ton of OC headroom throughout the GB203 GPU on the coronary heart of this card. If you happen to’re completely happy to spend a bit of time getting a secure overclock
Do not buy if…
❌ Inventory stays scarce: Till inventory ranges normalise I count on it is going to be mighty laborious to search out an MSRP card comparable to this MSI model whereas retailers sniff a chance to cost gouge.
If you happen to missed out on that RTX 5080 you initially had your coronary heart set on due to the shortage of inventory come launch day, then you definitely might need simply had a little bit of a fortunate escape. The brand new RTX 5070 Ti would possibly home a cut-down model of the identical GPU, but it surely’s not too far behind on uncooked body charge efficiency in any respect.
The MSRP card I have been testing for launch day is the MSI RTX 5070 Ti Ventus 3X. Regardless of that, it’s an overclocked card, although I’ve been assured it will hit that worth level within the UK not less than. For a way lengthy that can final, or what number of will really be obtainable, nevertheless, I’ve no such excellent news to share.
And I’ve very distinct worries about what this actual card is actually going to value in the true world.
Nonetheless, allow us to assume that is the bottom charge of an RTX 5070 Ti, as a result of I am a trusting type and that pleases me. The unique Nvidia whitepaper suggests the usual Enhance clock of the cardboard can be 2.45 GHz however, although these figures are at all times kinda worst case eventualities, I wasn’t anticipating this card to rock up with a 2.7 GHz common Enhance out of the field.
And but right here we’re, with a GPU able to excessive clock speeds, impressively low thermals, and being simply wanting the RTX 5080’s gaming efficiency by 14% at 1440p and 17% at 4K. Whenever you sit again and contemplate what that truly interprets to when it comes to body charges, you may see a card that is not often greater than 10 fps off the tempo.
Already, that seems like dangerous information for the RTX 5080, but it surely will get worse while you begin pushing these Enhance clocks even larger. From overclocking the RTX 5080 I knew the GB203 chip had a great deal of headroom obtainable, and so it has proved with the RTX 5070 Ti’s lower down model of that GPU, too.
With just a bit cavalier clock tweaking I am inside 10% of the RTX 5080 in any respect resolutions.
There’s something disgracefully smelly occurring with the RTX 5070 Ti pricing.
In addition to being round 25 – 30% faster than the outdated RTX 4070 Ti Tremendous, it is knocking on the door of the RTX Blackwell card above it, too. If you would like an older reference level, it is a GPU that hoses the RTX 3090, in gaming and productiveness phrases. Certain, sooner or later the additional 8 GB within the outdated $1,500 card will come into play, however solely in very VRAM restricted circumstances.
I do not know why the headroom has been left in there and never already exploited out of the field, however the overclocking efficiency nonetheless makes me much more impressed with this nonetheless relatively dear GPU. Of all of the high-end new Nvidia graphics playing cards—and at $749 it’s nonetheless very a lot that, even simply when it comes to base pricing alone—that is the one which I might suggest.
However sadly, I do have to complete my verdict with a fairly main caveat. Within the construct as much as launch it is turn into clear there’s something disgracefully smelly occurring with the RTX 5070 Ti pricing. The nominal $749 MSRP seems prefer it would possibly solely be paid passing lip service by retailers, with a bunch of listings already near the $1,000 mark.
We’re not immune within the UK, both, with one retailer telling me that it might be stocking the brand new playing cards beginning at £875 because the lowest worth, and different overclocked playing cards having their very own MSRP of £940.
At these types of costs the shine has effectively and really been scorched from the RTX 5070 Ti. If that is the place we land at for the standard retail worth—whether or not for weeks, months, or years—it will likely be an enormous disgrace. At its Nvidia MSRP it is already some huge cash, however the default efficiency, the Multi Body Gen characteristic, and the overclocking headroom all make it a good worth GPU.
But when producers and retailers are going to show scalper and cost a daft worth premium on prime then the RTX 5070 Ti is bound to have a rocky begin to life.
Nvidia RTX 5070 Ti: The specs
Picture 1 of 3
The RTX 5070 Ti is a direct step on from the RTX 4070 Ti Tremendous relatively than the unique RTX 4070 Ti, which is 100% a superb factor. Meaning, relatively than utilizing a decrease spec GPU from the beginning, Nvidia has opted to as an alternative stick with it utilizing only a cut-down model of the identical chip it is used for the RTX 5080.
That has a few key advantages, each for us and the inexperienced staff. For a begin it means we get a good reminiscence sub-system, with an aggregated 256-bit bus relatively than the 192-bit bus we’re set to get with the GB205 GPU going into the RTX 5070. Mixed with the identical 16 GB of GDDR7—albeit with a barely decrease clock velocity and rather less L2 cache—we have a really comparable setup to the RTX 5080 immediately.
This was how the RTX 4070 Ti Tremendous (oh, how I nonetheless hate that doubling up of suffixes…) was constructed, and I’m very happy Nvidia has caught to that. However as I say, it advantages the corporate, because it permits for failed GB203 chips, that might not have been used for RTX 5080 playing cards, to be repurposed for one thing else.
| Header Cell – Column 0 |
RTX 5070 Ti |
RTX 5080 |
RTX 4070 Ti Tremendous |
|---|---|---|---|
|
GPU |
GB203 |
GB203 |
AD103 |
|
CUDA cores |
8960 |
10752 |
8448 |
|
Enhance clock (GHz) |
2.45 |
2.62 |
2.61 |
|
Base clock (GHz) |
2.30 |
2.3 |
2.34 |
|
Tensor core TOPS (FP16) |
176 |
225 |
176 |
|
Ray tracing core TFLOPS |
133 |
171 |
102 |
|
Reminiscence |
16 GB GDDR7 |
16 GB GDDR7 |
16 GB GDDR6X |
|
Reminiscence bus width |
256-bit |
256-bit |
256-bit |
|
Reminiscence bandwidth (GB/s) |
896 |
960 |
672 |
|
Complete Graphics Energy (watts) |
300 |
360 |
285 |
|
Required system energy (PSU wattage) |
750 |
850 |
700 |
|
Energy connector |
1x 300 W PCIe Gen 5 OR 2x PCIe 8-pin adapter |
1x 450 W PCIe Gen 5 OR 3x PCIe 8-pin adapter |
1x 450 W PCIe Gen 5 OR 3x PCIe 8-pin adapter |
|
MSRP |
$749 |
$999 |
$799 |
Now, the TSMC 4N course of is identical because the Ada era, and as such should be fairly sturdy by now, and it is really a barely smaller GPU in contrast with the AD103 of the RTX 4080 and RTX 4070 Ti Tremendous, so yields must be good. Presumably that is why Nvidia was completely happy to make use of a full GB203 die for the RTX 5080s. However that does not go away it any wiggle room for chips that do not have a totally purposeful complement of 10752 CUDA cores.
Having the RTX 5070 Ti, with its 8960 cores, permits Nvidia to drop these chips that miss the complete mark right into a nonetheless very efficient graphics card.
The general energy ranges are similar to the earlier generations of RTX xx70 Ti collection playing cards, with the RTX 3070 Ti coming in with a 290 W TGP, the RTX 4070 Ti/Tremendous playing cards sporting a 285 W TGP, and this new RTX 5070 Ti driving a 300 W TGP.
There isn’t a Founders Version of the RTX 5070 Ti, so there’s nothing to be mentioned for Nvidia’s funky designs on this spherical of playing cards, but it surely offers the AIBs someplace to go together with their very own designs. This MSI RTX 5070 Ti Ventus 3X is essentially much like the RTX 4070 Ti Tremendous Ventus 3X we’ve got, with some slight variations within the design. It is rather less angular, which is to its credit score, however nonetheless retains the identical gray and black aesthetic and is available in a barely greater than dual-slot config.
Nvidia RTX 5070 Ti: The efficiency
Picture 1 of 2
The bottom efficiency of the RTX 5070 Ti is roughly according to what Nvidia mentioned it might be once we had been first taken by the brand new RTX Blackwell era of playing cards again on the CES Editor’s Day. There we had been informed to count on round a 20% gen-on-gen enhance over the RTX 4070 Ti Tremendous, and that is roughly the place we’re.
There may be one outlier in our testing suite, which is the Metro Exodus Enhanced Version. That is solely giving us an 11% efficiency bump, and is dragging down the general common, however the remainder of our check suite is exhibiting the identical 20% enhance.
That does then put it nearly instantly in the course of the RTX 4070 Ti Tremendous and the RTX 5080, being some 15-20% slower than the second-tier RTX Blackwell GPU. When it is nominally 25% cheaper—MSRP to MSRP—then that is not a foul return.
Nvidia desires us to consider the efficiency within the face of the RTX 30-series, provided that upgrading each era is fairly uncommon. However although you are one thing like a 50% efficiency bump over the RTX 3070 Ti, that was successfully one other tier under this card when it comes to its $599 pricing. Nonetheless, the truth that you may get the RTX 5070 Ti to run at round 25-30% sooner than the RTX 3090—the $1,500 flagship of the Ampere era—it is apparent there’s a wholesome enhance in gaming body charges leap-frogging a GPU gen.
You then even have the added advantages of Multi Body Era within the video games that assist it, that are most video games that at the moment assist common 2x Body Era. I am nonetheless mightily impressed with MFG as a characteristic—and it really makes this a wonderful 4K gaming card in these video games. I’ll say, that in one thing as taxing as Alan Wake 2, the place the enter body charge and PC latency aren’t nice, that issues begin to break down on the full 4x MFG setting. Dial that again to 3x and you are still getting an incredible stage of efficiency, however with fewer of the visible artifacts that appear to creep in.
That mentioned, with Cyberpunk 2077 on RT Overdrive at 4K, I am solely getting 15 fps as normal, and a brutal 181 ms of PCL. Throw DLSS High quality into the combination and at 4X MFG you may hit 110 fps, with simply 77 ms of latency. In a single participant recreation, that basically would not really feel like a lot in any respect.
And that stage of efficiency solely improves while you begin speaking about overclocking. In trendy GPU historical past, that hasn’t been one thing I’ve paid a lot consideration to in a assessment due to the truth that chips have been tuned to make the very best use of the headroom obtainable to them out of the field, so there was valuable little left for us nerdy tweakers. Not so with the RTX Blackwell era, and particularly the GB203 GPU.
I’ve simply booted up MSI Afterburner and recklessly pushed the GPU and reminiscence frequencies up.
After taking the RTX 5080 for a spin, I used to be excited to get my palms on the RTX 5070 Ti, and although this Ventus 3X would not have some kind of excessive triple-slot cooling array atop it I’ve nonetheless managed to get a ton of additional efficiency out of the GPU. Regardless of an preliminary bug that stopped me from altering the ability limits—which has since been mounted—I have been in a position to push this chip as much as +450 MHz stably, with one other +1000 MHz added onto the reminiscence clocks.
With these settings locked in, the hole between the RTX 5080 and the RTX 5070 Ti drops to single digits, percentage-wise, and the delta to the RTX 4070 Ti Tremendous grows to some 30%.
And it is not laborious to do both. I’ve not dug into any sorts of voltage tweaking or excessive cooling to get to these ranges both; I’ve simply booted up MSI Afterburner and recklessly pushed the GPU and reminiscence frequencies up. It falls over gaming at +475 MHz, and fully locks the moment you set +500 MHz, however in any other case it is secure with the marginally decrease settings in place.
And it does virtually nothing to the ability and thermal calls for of the chip, both. In brief, I do not see any motive why you should not simply set and neglect the overclock and stick with that efficiency stage. Clearly, the RTX 5080 will also be overclocked to hell and again, and which means it may possibly regain its lead, however the truth you may push the RTX 5070 Ti to shut to its default velocity continues to be tremendous spectacular.
Although I additionally do not see any motive why both the GB203 playing cards weren’t given 3 GHz+ increase clocks out the field.
On the creator aspect, the RTX 5080 retains extra of an edge, although as they’re each sporting primarily the identical stage of VRAM there nonetheless is not so much between them. It’s fascinating, nevertheless, that within the PugetBench for DaVinci Resolve benchmark there’s virtually nothing between them on the AI rating, regardless of the precise AI picture era exams nonetheless closely favouring the RTX 5080.
Price noting additionally that even the place this RTX 5070 Ti Ventus 3X is drawing extra energy than its forebear, it is solely by a modest quantity, and although it is received successfully the identical sized cooling shroud, it really runs a bit of cooler. That is most likely as a result of, although it would not have as many heatpipes inside it, they appear a superb chunk fatter.
PC Gamer check rig
CPU: AMD Ryzen 7 9800X3D | Motherboard: Gigabyte X870E Aorus Grasp | RAM: G.Talent 32 GB DDR5-6000 CAS 30 | Cooler: Corsair H170i Elite Capellix | SSD: 2 TB Essential T700 | PSU: Seasonic Prime TX 1600W | Case: DimasTech Mini V2
Nvidia RTX 5070 Ti: The Evaluation
Picture 1 of 3
A probably nice graphics card that appears like it is going to undergo from being priced into oblivion.
The RTX 5070 Ti is a probably nice graphics card that appears like it is going to undergo from being priced into oblivion. The problem is that whereas the MSRP from Nvidia seems nice, with a $750 price ticket representing a $50 lower from the $799 of the RTX 4070 Ti/Tremendous playing cards, it appears each producers and retailers are aiming to money in on the shortage and demand of the brand new RTX 50-series GPUs.
The RTX 5090 and RTX 5080 vanished from inventory inside 5 minutes of launch, and although there could also be larger volumes of the RTX 5070 Ti within the channel given the later launch date, they’re more likely to disappear rapidly, too. Particularly as there aren’t any different options round. Each AMD and Nvidia wound up manufacturing of its last-gen graphics playing cards in direction of the top of final 12 months, and retailers bought by their remaining inventory in anticipation of latest playing cards from the inexperienced and pink aspect of the graphics divide.
This implies GPU makers are eyeing up a chance to cost gouge with impunity. They go in figuring out that even when it’d put the vast majority of avid gamers off with ludicrous pricing, there can be sufficient individuals keen to spend large simply to get a brand new GPU that these few will deal with the restricted inventory obtainable at launch.
Producers then get the added bonus of pretending an enormous worth lower down nearer to MSRP afterward when the inventory ranges begin to normalise. Win, win for them, an enormous ol’ L for PC avid gamers.
The MSI RTX 5070 Ti Ventus 3X I’ve right here for testing is theoretically an MSRP model, so sayeth Nvidia, however I might be stunned in the event you see it on sale within the US for $750. And I do know for positive there are comparable playing cards, with comparable efficiency, which might be going to go on sale for nearer to $1,000, the unique MSRP of the RTX 5080.
I might now count on most RTX 5070 Ti playing cards are going to be priced up far nearer to the RTX 5080 than the set Nvidia MSRP for this GPU, not less than over the subsequent few weeks and months. And in that occasion, the very fact there’s little when it comes to efficiency between the 2 playing cards—one of many major causes I’ve felt so constructive in regards to the RTX 5070 Ti—means much less and fewer because the distinction in pricing vanishes, too.
