GeForce RTX 4090 Leaves Plenty of Room for a Future RTX 4090 Ti Flagship
When you purchase through links on our site, we may earn an affiliate commission. Here’s how it works.
(Image credit: Nvidia)
Nvidia’s GeForce RTX 4090 might look incredibly strong, and will certainly rank as the fastest option on our list of the best graphics cards when it debuts (at least until AMD’s RDNA 3 GPUs arrive), but the shaved down AD102 die in the RTX 4090 isn’t close to showing off the full potential of AD102 with all of its cores and cache enabled. This combined with additional enhancements could hint at a future RTX 4090 Ti that will be much faster — and perhaps even more expensive.
The specs for the Nvidia RTX 40-series and Ada Lovelace GPUs, but those only show the announced and rumored cards. Nvidia’s full AD102 die comes equipped with 144 SMs, 18,432 CUDA cores, 96MB of L2 cache, and 192 ROPs. This translates to 12% more CUDA cores and a whopping 33% more L2 cache capacity compared to the RTX 4090 we have today. The fully enabled AD102 die also packs 9% more ROPS and 12% more Texture Mapping Units as well, thanks to the additional SMs.
But that’s not all that could be done for the future 4090 Ti. Micron has new 24Gbps GDDR6X memory modules in the works, another 14% boost over the RTX 4090’s 21Gbps modules, and still faster than the RTX 4080 16GB’s 22.4 Gbps modules that Nvidia claims are the fastest in the world right now. That would push the hypothetical (but very likely) RTX 4090 Ti up to 1152 MB/s of bandwidth.
But faster memory would come with higher power consumption, and we suspect that Nvidia is seriously holding back AD102’s full clock speed and power potential as well. All those rumors of 600W RTX 40-series graphics cards? We know Nvidia has successfully overclocked RTX 4090 to more than 3.0GHz, and that would definitely push up power use.
It looks like the Ada architecture and TSMC’s 4N process have plenty of headroom remaining beyond the RTX 4090’s 2520 MHz boost frequency.
Once the process matures a bit more, and if Nvidia is willing to increase the power limits, we wouldn’t be surprised to see a RTX 4090 Ti clock at closer to 2800 MHz.
The theoretical performance of AD102 with all these bells and whistles enabled could reach a whopping 103 teraflops in FP32 workloads, and 826 teraflops in FP16 workloads with the Tensor cores, and 1652 teraflops with the Tensor cores in FP8 mode. That would be a huge 25% performance jump in comparison to the RTX 4090.
These gains would only be realized in GPU limited scenarios, of course, so probably not 1080p or 1440p gaming. Heavy compute applications would also likely benefit. The combination of more L2 cache capacity, additional GDDR6X bandwidth, and more cores and clocks could result in tangible improvements.
Swipe to scroll horizontally
| Row 0 – Cell 0 | RTX 4090 Ti (Full AD102) | RTX 4090 | RTX 3090 Ti |
| Process | TSMC 4N | TSMC 4N | Samsung 8N |
| Transistors | 76. 3B |
76.3B | 28.3 |
| SMs | 144 | 128 | 84 |
| GPU Cores | 18432 | 16384 | 10752 |
| Tensor Cores | 576 | 512 | 336 |
| Ray Tracing Cores | 144 | 128 | 84 |
| Boost Clock | 2800MHz??? | 2520MHz | 1860MHz |
| VRAM Speed | 24 Gbps? | 21 Gbps | 21Gbps |
| VRAM | 24GB | 24GB | 24GB |
| Bus Width | 384 | 384 | 384 |
| Memory Bandwidth | 1152GB/s | 1008GB/s | 1008GB/s |
| L2 Cache Capacity | 96MB | 72MB | None |
| ROPs | 192 | 176 | 112 |
| TMU | 576 | 512 | 336 |
| TFLOPS FP32 | 103. 2 |
82.6 | 40 |
| TFLOPS FP16 | 826 | 661 | N/A |
| TDP | 600W?? | 450W | 450W |
It appears Nvidia has a lot of performance headroom remaining with its GA102 die, with the potential to create a RTX 4090 Ti that could theoretically smoke the RTX 4090. It would certainly cost a lot more money, and consume way more power than a RTX 4090, but it can be done.
All of this will depend on how hard Nvidia wants to push its GA102 die, and that will almost certainly depend on how close AMD can come to matching Nvidia’s performance with the upcoming RDNA 3 chips. Yields on fully functional AD102 GPUs would also play a role, though it’s doubtful these would be high volume parts.
Nvidia could add some or all of these enhancements to an RTX 4090 Ti any time it feels the need.
We didn’t get the RTX 3090 Ti until 18 months after the RTX 3090 debut, but there were a lot of compounding factors in play. More likely is we’ll see a 2023 refresh of the RTX 40-series some time around nine months to 12 months after the initial salvo.
There’s also the rare chance Nvidia could skip the RTX 4090 Ti completely in favor of a new Titan variant, but we doubt that will be the case. Titan cards tend to cut into the lucrative RTX A-series professional card profits too much.
Join the experts who read Tom’s Hardware for the inside track on enthusiast PC tech news — and have for over 25 years. We’ll send breaking news and in-depth reviews of CPUs, GPUs, AI, maker hardware and more straight to your inbox.
Contact me with news and offers from other Future brandsReceive email from us on behalf of our trusted partners or sponsors
Aaron Klotz is a freelance writer for Tom’s Hardware US, covering news topics related to computer hardware such as CPUs, and graphics cards.
Best graphics cards for PC gaming 2023
Updated
We’ll help you find the best graphics card to fit your needs.
By Brad Chacos
Executive editor, PCWorld Feb 23, 2023 7:30 am PST
Image: Rob Schultz/IDG
Most people who are in the market for a new graphics card have one primary question in mind: Which card will give me the most bang for my buck? Obviously, the answer will vary depending on your budget. Beyond that, there are a number of factors to consider: Raw performance is important, but so are things like noise, the driver experience, and supplemental software. And do you want to pay a premium to experience the gorgeous, but performance-intensive visuals possible with real-time ray tracing?
Let us make it easy for you. We’ve tested nearly every major GPU that’s hit the streets over the past couple of years, from $100 budget cards to $2,000 luxury models. Nvidia, AMD, Intel Arc, we’ve benchmarked them all.
Our knowledge has been distilled into this article—a buying guide with recommendations on which graphics card to buy, no matter what sort of experience you’re looking for.
Note: There are customized versions of every graphics card from a host of vendors. For example, you can buy different GeForce GTX 3080 models from EVGA, Asus, MSI, and Zotac, among others.
We’ve linked to our complete review for each recommendation, but the buying links lead to models that hew closely to each graphics card’s MSRP. Spending extra can get you hefty out-of-the-box overclocks, beefier cooling systems, and more. Check out our “What to look for in a custom card” section below for tips on how to choose a customized card that’s right for you.
The best graphics cards for PC gaming
AMD Radeon RX 6500 XT – Best graphics card under $200
Prices may be relaxing, but currently, the much-maligned Radeon RX 6500 XT is still the only semi-reasonable sub-$200 option around.
The Radeon RX 6500 XT is less appealing thanks to its nerfed memory, PCIe lanes, and limited ports, not to mention lower performance, but you can often find them going for around $170 on the streets these days. Those hardware limitations mean you’ll need to stick to Medium or High graphics settings at 1080p resolution in modern games in order to achieve playable frame rates, but if you do that, you’ll enjoy the experience. This card will get you by, but if it’s possible, save a bit longer and splurge on one of our favorite 1080p graphics cards for about $250 instead.
Read our full
Radeon RX 6500 XT review
AMD Radeon RX 6600 – Best 1080p graphics card
AMD’s Radeon RX 6600 and Nvidia’s rival GeForce RTX 3060 both ostensibly carry the same $329 MSRP, but on the streets, there’s a much wider gap. You can find the Radeon 6600 regularly going for $200 to $250, while the cheapest RTX 3060 usually cost around $400.
Those are both steep entry costs for 1080p gaming—at least compared to the GPUs of yesteryear—but with 8GB of fast GDDR6 memory, insanely good power efficiency, and AMD’s Radeon Super Resolution in tow, the Radeon RX 6600 is a great graphics card for people looking to game at 1080p resolution at 60fps or higher without compromising on visual fidelity. (Or breaking the bank.) It lags in ray tracing performance, however.
Read our full
Radeon RX 6600 Swft 210 review
Intel Arc A750 – Best 1080p graphics card for ray tracing
Intel’s debut Arc graphics cards best even Nvidia’s vaunted RTX 30-series graphics cards at ray tracing in this price class, and the newly $250 Arc A750 does it for about $150 less than the popular GeForce RTX 3060. Why compare the A750 against that $400ish card rather than the $300ish RTX 3050? Because if you’re playing newer games running on modern DirectX 12 or Vulkan APIs—most triple-A games, in other words—the Arc A750 meets or beats the RTX 3060 there, too.
At launch, Intel’s GPUs were plagued by bugs and odd performance. Since then, the company worked diligently to improve its drivers at a torrid pace, unleashing exceptional performance improvements in DX9 games (read: esports and classics) while also squashing the most worrisome bugs. Performance remains hit-or-miss on older DirectX 11 titles, which is why the Radeon RX 6600 remains our top pick overall, but its definitely good enough—and still getting better. But if your focus is mostly on ray tracing and triple-A games, the Intel Arc A750 is a much better option at the same $250 price point as the 6600.
Read our full
Intel Arc A750 Limited Edition review
AMD Radeon RX 6700 XT – Best 1440p graphics card
In a sane world, Nvidia’s GeForce RTX 3060 Ti would dominate 1440p gaming at its $400 MSRP. It’s that good, and it offers superior ray tracing performance to AMD’s Radeon rivals. But we still aren’t living in a sane world, and the RTX 3060 Ti is going for $500+ on the streets, and often $550 to $600.
Nvidia’s RTX 3070, ostensibly $500, goes for $650 to $700 online.
Get AMD’s Radeon RX 6700 XT instead, for a whole lot less. It’s plenty fast for 1440p gaming at 60fps+ without compromise, while its beefy 12GB of GDDR6 memory provides plenty of headroom for flipping on all the most intense graphical features. The one downside? AMD’s card is only capable of playing ray-traced games at 1080p resolution unless you activate Radeon Super Resolution, or FSR 1 or 2 in games that support it. One the flip side, the Radeon RX 6700 XT can take advantage of AMD’s awesome performance-boosting Smart Access Memory feature if you’re running a modern Ryzen system that supports it.
Read our full
Radeon RX 6700 XT review
Nvidia GeForce RTX 4070 Ti – Best high-end 1440p graphics card
Paying $800 or more for a 1440p graphics card is ludicrous. Our review called Nvidia’s GeForce RTX 4070 Ti “painfully overpriced.
” But if you want exceptional performance at 1440p gaming—and only 1440p gaming, as Nvidia hobbled this $800 GPU to run slower at 4K—price be damned, the 12GB RTX 4070 Ti is a great option. It’s ultra-fast, ultra-efficient, excellent at ray tracing, and supports Nvidia’s killer new DLSS 3.0 Frame Gen technology. If you want still-solid 1440p gaming with the potential for dabbling in 4K, with less efficiency (but more memory) at a lower price, also consider the next option instead.
Read our full
GeForce RTX 4070 Ti review
AMD Radeon RX 6800 XT – Best affordable 4K graphics card
The new generation is here, so it’s time for discounts on last-gen flagships—at least on AMD’s side. While older Nvidia’s RTX 30-series graphics cards remain priced near (or even above) MSRP over two years after launch, Radeon is cutting prices. And thanks to that, you can often score the ferocious Radeon RX 6800 XT for under $600 on the street.
It cost $800 at launch and goes blow-for-blow with Nvidia’s popular RTX 3080—but now for hundreds less. And AMD supplied the Radeon RX 6800 XT with an ample 16GB of GDDR6 memory, which holds up a lot better for modern 4K gaming than the RTX 3080’s mere 10GB.
If you want to keep a 4K/60 monitor fed with all your game’s eye candy turned on—except ray tracing, which this card struggles with—the Radeon RX 6800 XT is a fantastic buy. So are the Radeon RX 6800 and step-up 6900 XT, which can also be found at steep discounts these days.
AMD Radeon RX 7900 XTX – Best 4K graphics card
If you’ve got a 4K monitor and want to put all those pixels to work, the RTX 3080 and RTX 3070 are decent cheaper options—but they’re last-generation GPUs with still-inflated price tags that simply don’t make sense now that a new breed of graphics cards are here. Enter AMD’s flagship Radeon RX 7900 XTX. This $999 can keep a 4K/60 monitor maxed out without breaking a sweat, and often flirts with maxing out 4K monitors with higher refresh rates.
Yes, that’s with all the graphics options enabled—except ray tracing. Long a losing battle for AMD, the Radeon RX 7900 XTX remains firmly behind the ray tracing prowess of Nvidia’s newer GeForce RTX 40-series GPUs, but still slings rays slightly better than Nvidia’s last-gen flagship, the monstrous RTX 3090 Ti. That’s pretty damned good.
If you do video editing or run Minecraft with tons of mods enabled, you’ll also appreciate the Radeon RX 7900 XTX’s ample 24GB memory capacity. The $800 GeForce RTX 4070 TI, $900 Radeon RX 7900 XT, and $1,200 GeForce RTX 4080 are all in a similar price class to the 7900 XTX, but offer atrocious value for the performance delivered. Get this, get a last-gen card…or splurge on our next option.
Read our full
Radeon RX 7900 XTX review
Nvidia GeForce RTX 4090 – Best high-end 4K graphics card
Graphics cards that cost $1,000 didn’t used to exist, but now they’re commonplace.
All of them offer a compelling 4K gaming experience. But if you want peak performance no matter the price, you’ll be spectacularly pleased with the GeForce RTX 4090 Founders Edition. This is the first graphics card capable of maxing out a 120Hz 4K monitor in many modern games—a monumental achievement. Its rivals can kept a 60Hz 4K monitor fed, but not a high refresh rate display. The GeForce RTX 4090 embarrasses all previous GPU contenders in all games, full stop—it’s so fast that games can become CPU bottlenecked even at 4K/Ultra when using this GPU. You’ll want a new Radeon 9 7950X or Intel Core i9-13900K to truly take advantage of it.
Beyond the raw speed, the RTX 4090 also offers an ample 24GB of GDDR6 memory for content creators, best-in-class ray tracing chops, and new DLSS 3 technology that uses AI to improve frame rates substantially—think twofold or even more, though responsiveness can take a hit in some games. The RTX 4090 is an absolutely sublime GPU for peak-performance gaming.
Read our full
GeForce RTX 4090 Founders Edition review
How we test graphics cards
We test graphics cards on a dedicated test system used only for this purpose, with minimal extra software involved. That ensures that any performance changes we see are generated solely by the graphics card being tested and new GPU drivers, without the variability of other hardware or software changes. Here is the configuration of our current testbed:
- AMD Ryzen 5900X, stock settings
- AMD Wraith Max cooler
- MSI Godlike X570 motherboard
- 32GB G.Skill Trident Z Neo DDR4 3800 memory
- EVGA 1200W SuperNova P2 power supply
- 2x 1TB SK Hynix Gold S31 SSD
As far as games go, we use a fixed set of games to test every graphics card that comes out in a given generation, and update the suite when a new generation of GPUs is introduced. We test a variety of games spanning most major game types (tactics, racing, FPS, etc.
), engines (Unreal Engine, Unity, Anvil, etc.) and underlying graphics APIs (DirectX 11, DX12, Vulkan). We use the built-in benchmarks for each game, but only after validating the accuracy of the results by running the benchmarks and comparing the results to performance witnessed by third-party GPU measurement tools like OCAT. Each game is tested at least three times per resolution, generating an average from those runs, with additional tests run if we encounter any hiccups. We may also perform additional testing with tools like OCAT if any performance oddities are noticed. Power draw is measured on a whole-system basis, listing both idle and fully stressed states as measured via a Watts Up meter that the system is plugged into.
What to look for in a custom graphics card
If you want to shop beyond the scope of our picks, know that finding the right graphics card can be tricky. Various vendors offer customized versions of every GPU. For example, you can buy different Radeon RX 6700 XT models from Sapphire, XFX, Asus, MSI, and PowerColor.
To help narrow down the options and find the right card for you, you should consider the following things when doing your research:
Overclocks: Higher-priced custom models are often overclocked out-of-the-box to varying degrees, which can lead to higher performance. Most modern custom cards offer the same essential level of performance,however.
Cooling solutions: Many graphics cards are available with custom coolers that lower temperatures and fan noise. The vast majority perform well. Liquid-cooled graphics cards run even cooler, but require extra room inside your case for the tubing and radiator. Avoid graphics cards with single-fan, blower-style cooling systems if you can help it, unless you have a small-form-factor PC or plan on using custom water-cooling blocks.
Size: Many graphics cards are of a similar size, but longer and shorter models of many GPUs exist. High-end graphics cards are starting to sport especially massive custom cooling solutions to tame their enthusiast-class GPUs.
Double-check that your chosen graphics card will fit in your case before you buy.
Compatibility: Not all hardware supports a wide range of connectivity options. Higher-end graphics cards may lack DVI ports, while lower-end monitors may lack DisplayPorts. Only the most modern Radeon and GeForce graphics cards support HDMI 2.1 outputs. Ensure your graphics card and monitor can connect to each other. Likewise, make sure your power supply meets the recommended wattage for the graphics card you choose.
Real-time ray tracing, FSR, and DLSS: AMD’s Radeon RX 6000-series graphics cards and all of Nvidia’s RTX offerings can play games with real-time ray tracing effects active. Nvidia’s RTX 30-series GPUs hold a massive advantage over everything else though, propelled even further by dedicated tensor cores for processing machine learning tasks such as Deep Learning Super Sampling, which uses AI to speed up the performance of your games with minimal hit to visual fidelity.
GeForce RTX 20-series GPUs also support DLSS, while AMD’s rival FSR 2.0 and Radeon Super Resolution technologies are gaining traction by the day.
Author: Brad Chacos, Executive editor
Brad Chacos spends his days digging through desktop PCs and tweeting too much. He specializes in graphics cards and gaming, but covers everything from security to Windows tips and all manner of PC hardware.
2023 PC gaming graphics cards
NVIDIA or AMD, GTX or RTX? Buy now or wait for the next generation? Questions like these haunt you when it’s time for an upgrade. To help you solve at least some of them, we have compiled a list of the best graphics cards at the moment.
NVIDIA
NVIDIA GeForce RTX 3090 Ti
GPU – GA102 | CUDA Cores – 10,752 | Memory – 24GB GDDR6X | Bus – 384-bit
Pros :
- Ultimate Power
- Large memory
- Ideal for 4K resolution
Cons :
- Out of reach for most gamers
- Uses a lot of energy
The
GeForce RTX 3090 Ti was designed for enthusiasts who don’t accept compromise and are willing to pay for the pleasure of owning the world’s most powerful graphics card.
Indeed, it has an advanced graphics processor, and the list of supported technologies takes up a whole page. The graphics card will easily handle any modern games at 4K resolution with ray tracing enabled (especially if you use the DLSS function). Among the shortcomings of the video card – prohibitively high price, indefatigable power consumption and large dimensions.
NVIDIA GeForce RTX 3090
GPU – GA102 | CUDA cores – 10,496 | Memory – 24GB GDDR6X | Bus – 384-bit
Pros :
- Very powerful
- Large memory
- Good for both gaming and professional use
Cons :
- Very expensive
The Nvidia GeForce RTX 3090 costs $500 less than the GeForce RTX 3090 Ti, while not losing much power to the older model. It’s a great graphics card for 4K and 1440p gaming as well as professional tasks like video editing and 3D modeling.
However, it is still very expensive, especially considering that it is not easy to find a video card at the suggested retail price. The adapter consumes 100 watts less power than the 3090 Ti, which will significantly offload your power supply. And take care of the presence of a spacious case that can accommodate an overall three-slot video card.
NVIDIA GeForce RTX 3080 Ti
GPU – GA102 | CUDA Cores – 10,240 | Memory – 12GB GDDR6X | Bus – 384-bit
Pros :
- Support for modern rendering and visualization technologies
Cons :
- Back-to-back memory for 4K games
- 12-pin power connector included
The
3080 Ti is designed with gamers in mind, so it supports all the latest technologies for an unrivaled gaming and streaming experience. Among them are DLSS, and ray tracing, and Nvidia Reflex, and an encoder for better image transmission through streaming, and Resizable BAR, and much more.
The graphics card is powerful enough to run games at maximum settings in 4K resolution and with DLSS: it provides more than 60 fps in all modern AAA titles. You can count on this graphics adapter to provide you with high performance for years to come.
NVIDIA GeForce RTX 3080
GPU – GA102 | CUDA Cores – 8,960 | Memory – 10/12GB GDDR6X | Bus – 320/384-bit
Pros :
- More powerful than RTX 2080 Ti
- Supports modern technologies
- Good for 1440p and 4K
Cons :
- Not enough memory to play in high resolution
- 12-pin power connector
The GeForce RTX 3080 outperforms the RTX 2080 Ti by approximately 15%, while the new graphics card packs more features than the previous generation.
You can still play 4K with this video card and there is no game that it can’t handle at 1440p, but with ray tracing enabled, you’ll have to make compromises in the form of lower settings.
And using the newfangled 12-pin power connector will force you to look for an appropriate adapter or upgrade the power supply.
NVIDIA GeForce RTX 3070 Ti
GPU – GA104 | CUDA Cores – 6,144 | Memory – 8GB GDDR6X | Bus – 256-bit
Pros
Nvidia developed the GeForce RTX 3070 Ti as a “response” to its closest competitor, the AMD RX 6800. The answer was controversial: the model outperforms its rival in 4K resolution, but is slightly inferior in 1440p and 1080p. However, with the activation of ray tracing and DLSS, Nvidia’s adapter wins a clear victory.
This is a good graphics card if you don’t plan on playing at 4K resolution, where it’s getting too small, and by that we mean making trade-offs in the form of lower settings or frame rates. On the other hand, we would recommend paying a little extra and getting an RTX 3080.
NVIDIA GeForce RTX 3070
GPU – GA104 | CUDA Cores – 5,888 | Memory – 8GB GDDR6 | Bus – 256-bit
Pros
Cons :
- 12-pin connector
- Weak for 4K
The
RTX 3070 is great if you’re not going for max settings and high resolutions.
With this graphics card, you can comfortably play at 1440p, and it supports all modern Nvidia technologies, including RTX, DLSS, Resizable BAR and so on. In terms of performance, it compares with the RTX 2080 Ti, the flagship of the previous generation, which is a very good result for a middle-class video card.
At the same time, this model is smaller than the older representatives of its generation, less noisy and more energy efficient. The disadvantages include the presence of a 12-pin power connector, although this should not be a big problem, given that the manufacturer puts in the kit an adapter for power supplies that are not equipped with the appropriate connectors.
NVIDIA GeForce RTX 2080 TI
GPU – TU102 | CUDA Cores – 4,352 | Memory – 11GB GDDR6 | Bus – 352-bit
Pros :
- Serious Gaming – 4K/60fps
- Faster than $3,000 Titan V
- All the goodies of ray tracing and AI
Cons :
- Conservative price tag
Here it is, the true nextgen.
GPU based on the Turing architecture, ultra-modern and unsurpassed “king of graphics”. While the RTX 2080 and 2070 Super are able to match the performance of the GTX 1080 Ti, this card leaves them far behind, giving you 4K/60fps in all but the most demanding games today.
No matter how you feel about newfangled (for PC gaming) ray tracing technology, there’s no getting away from it in the future, and the RTX 2080 Ti is ready for it. With the release of PS5 and Xbox Series X consoles that support ray tracing, there are more and more games with appropriate graphic effects, and sooner or later you will have to upgrade.
NVIDIA GeForce RTX 3060 Ti
GPU – GA104 | CUDA Cores – 4,864 | Memory – 8GB GDDR6 | Bus – 256-bit
Pros :
- Ideal for playing at 1440p and 1080p
- Quiet
- Energy efficient
- Relatively affordable
GeForce RTX 3060 Ti could become a new “people’s” video card, if not for the shortage and high prices for devices in this segment of computer technology.
However, owners of the model can count on excellent performance at 1440p and 1080p, and for 1080p you don’t even need to use DLSS when ray tracing is activated – the video card can handle it.
Unlike the “monsters” of energy consumption in the face of older models, 3060 Ti is energy efficient and quiet. In terms of performance, it can be compared with the more expensive RTX 2080 SUPER, and we would recommend buying this video card to those who have finally decided to switch from a Pascal family video card to something new.
NVIDIA GeForce RTX 2080 SUPER
GPU – TU104 | CUDA Cores – 2,944 | Memory – 8GB GDDR6 | Bus – 256-bit
Pros :
- Excellent 4K gaming
- RTX Turing architecture
Cons
The days of the RTX 2080 were numbered as soon as Nvidia admitted to releasing a Super version soon. The new graphics card outperforms the GTX 1080 Ti with more cores and faster memory. And yet, the latter has the best performance with tracing the best at the turn of up to $1000. GPU – TU104 | CUDA Cores – 2,560 | Memory – 8GB GDDR6 | Bus – 256-bit Pros : Cons : The RTX 2070 Super was born to win with the new Navi chips. As AMD took aim at the old RTX 2070, Nvidia released an upgraded version that is cheaper and slightly inferior to the 2080. It’s fast. Really fast. Impressive 4K can be squeezed out of this card, far ahead of Pascal’s best result on the GTX 1080 Ti. GPU – GA106 | CUDA Cores – 3,584 | Memory – 12GB GDDR6 | Bus – 192-bit Pros : Cons : The GeForce RTX 3060 boasts a low suggested retail price, as well as good performance at 1080p and 1440p, combined with a whopping 12 GB of video memory. The latter, however, is debatable, since this amount is redundant for playing at resolutions lower than 4K. Although, on the other hand, there is never enough memory. The graphics card allows you to play with ray tracing, provided you’re willing to compromise and don’t mind turning on DLSS – without intelligent anti-aliasing, the RTX 3060 can’t do much with RT enabled. GPU – TU106 | CUDA Cores – 2.176 | Memory – 8GB GDDR6 | Tire – 256-bit Pros : Cons : The RTX 2060 is the bare minimum required to enter the magical world of ray tracing (even though the technology isn’t perfect yet, RTX games are truly transforming). More importantly, thanks to DLSS 2.0, you can play games at 1440p resolution with no fps loss and with maximum graphics settings with this video card. If you choose between RTX 2060 and RTX 2060 SUPER, then we would recommend the second one: it has better overclocking potential, more memory and higher frequency, which means it will stay relevant longer compared to the younger version. GPU – GA106 | CUDA Cores – 2,560 | Memory – 8GB GDDR6 | Bus – 128-bit Pros : Cons : GeForce RTX 3050 is a budget option for a gaming system. This model is distinguished from similar performance video cards of the middle segment from the previous generation by supporting a larger number of current technologies. Cool and quiet, the graphics card is able to deliver acceptable frame rates at 1080p. However, it is too weak for ray tracing games, although the available tensor cores do not feel superfluous – they can be used to activate DLSS and the subsequent increase in fps. GPU – TU116 | CUDA Cores – 1,536 | Memory – 6GB GDDR6 | Bus – 192-bit Pros : Nvidia’s consumer graphics card takes advantage of the shader benefits of the Turing 20xx series architecture, but lacks the ray tracing and intelligent processing that sets the RTX line apart from other GPUs. Overclocked cards like the STRIX GTX 1660 Ti are too expensive, closer to the RTX 2060 and RX 5700, but the GTX 1660 Ti has no rivals in RRP. GPU – TU116 | CUDA Cores – 1.408 | Memory – 6GB GDDR5 | tire – 192-bit Pros : Cons : Since the introduction of Polaris GPUs in the sub-$300 segment, the GTX 1660 has taken a turn for the worse. The Nvidia card still has a slight performance advantage, which gives it the lead, offsetting the cost difference. When we spend money, we expect to get the highest possible frame rate for them, and green GPUs still meet our needs. In addition, they are equipped with new technologies Integer Scaling (scaling without blurring) and Low Latency Mode (low latency mode), which AMD graphics cards lack. GPU – TU116 | CUDA Cores – 1,280 | Memory – 4GB GDDR6 | Bus – 128-bit Pros Cons
GeForce GTX 1650 Super is enough for games at 1080p or lower. If you’re willing to lower your graphics settings, you’ll be able to get high fps, and the Turing architecture’s support for DLSS technology will further improve performance. At the same time, 4 GB of video memory is frankly not enough for modern AAA games. This will have to be taken into account if you want to download a pack of high-quality textures for your favorite title. GPU – GP107 | CUDA Cores – 768 | Memory – 4GB GDDR5 | Bus – 128-bit Pros : Cons
If the RTX 2060 is the bare minimum for ray-traced gaming, then the GTX 1050 Ti is the bare minimum for gaming. Of course, you will have to make compromises with it: it is suitable for gaming at a resolution not higher than 1080p, and it is better to forget about the maximum graphics settings. On the other hand, it is very cheap (although it could be even cheaper) and will run all modern video games – even those from the AAA segment. However, it should be noted that with the release of a new generation of consoles, its performance will no longer be enough. GPU – Navi 21 | RDNA Cores – 5,120 | Memory – 16GB GDDR6 | Bus – 256-bit Pros0025 Cons : The AMD Radeon RX 6900 XT is a direct competitor to the Nvidia GeForce RTX 3080 Ti, only losing slightly in power. With this graphics adapter, AMD is returning to the top graphics card segment for the first time since the release of the R9 Fury X. In terms of pure performance, this model breathes down the back of the Nvidia GeForce RTX 3090, but in real-world performance, things are a little more complicated, since Nvidia has an trump card in in the form of DLSS technology. On the other hand, the “red” video card is cheaper than the “green” competitors, which can be a weighty argument in its favor. GPU – Navi 21 | RDNA cores – 4,608 | Memory – 16GB GDDR6 | Bus – 256-bit Pros : Cons : The AMD Radeon RX 6800 XT takes on the Nvidia GeForce RTX 3080 by being roughly on par with the competition but with more memory, which is great for 4K and 1440p gaming. All this is true on one condition: you are ready to abandon ray tracing in games. Alas, in terms of RT support, AMD video cards lose to Nvidia solutions, so when you enable the appropriate effects, you will encounter a strong drop in frame rate. It is worth considering this nuance. GPU – Navi 22 | RDNA Cores – 2,560 | Memory – 12GB GDDR6 | Bus – 192-bit Pros : Cons : AMD promises that the Radeon RX 6700 XT is able to outperform not only the GeForce RTX 3060 Ti, but even the RTX 3070, at a more modest price tag than its competitors from Nvidia. This is a great choice for 1440p games. The video card is compact, quiet and “cold”. It has 12 GB of video memory, which allows you to use high-resolution textures, and reasonable power consumption. However, in terms of performance per dollar, this is not the most economical solution, so we recommend carefully weighing the pros and cons. GPU – Navi 23 | RDNA Cores – 2,048 | Memory – 8GB GDDR6 | Bus – 128-bit Pros Cons : The AMD Radeon RX 6600 XT is a smart alternative to GPUs like the Nvidia GeForce RTX 3060/Ti that delivers good gaming performance at 1080p. Traditionally for AMD video cards, this model does not cope well with ray-traced games due to the fact that competitors have DLSS support. AMD has an answer for this in the form of FSR (FidelityFX Super Resolution), but this technology is not as well established as DLSS, and is supported by fewer games. GPU – Navi 10 | RDNA Cores – 2,560 | Memory – 8GB GDDR6 | Bus – 256-bit Pros : Cons : With aggressive pricing, AMD has turned out to be a real monster that can not only make the RTX 2070 Super tense, but also save money at the same time. The best part about this is that for the money, the RT 5700 XT almost takes us to the promised land of high-end GPUs, challenging the “green” gang. Still, it (in its reference form) would not be so noisy and burning. GPU – Navi 24 | RDNA Cores – 1.024 | Memory – 4GB GDDR6 | Bus – 64-bit Pros : Cons : The main disadvantage of the AMD Radeon RX 6500 XT is its small amount of memory: with 4 GB, you can forget about decent gaming performance at resolutions above 1080p. Of course, one should not even dream of ray tracing in this case. Add to this the 64-bit bus, which also affects the efficiency of the video card, and you have an option that should be settled only as a last resort. AMD, however, takes an affordable price tag and the lack of alternatives in the budget segment. For the most part, this is a plug card, but for lack of a better one, it will do. GPU – Navi 10 | RDNA Cores – 2.304 | Memory – 8GB GDDR6 | Bus – 256-bit Pros : Cons : The RX 5700, which runs on the same chip, but with cut frequencies, immediately follows its XT companion. By the way, this also affects overclocking: even in capable hands, it will still be inferior to its older sister. And it’s sad, because the RX 5700 is a great powerful graphics card with a moderate price and performance that can put Nvidia to shame in its price segment. GPU – Navi 10 XLE | RDNA Cores – 2. Pros : The AMD Radeon RX 5600 XT performs well at 1080p and slightly worse at 1440p, but at 4K resolution it will show modest results. Its main advantage is the new RDNA architecture based on the 7nm process technology, due to which gamers can expect impressive results in DX12 applications, where the video card confidently outperforms its competitor from the “green” camp – NVIDIA GeForce RTX 2060. that it would be nice to add a little and get a more powerful RX 5700 XT. GPU – Navi 14 | RDNA cores – 1.408 | Memory – 4/8GB GDDR6 | Bus – 128-bit Pros : Cons : For 1080p gaming, the AMD Radeon RX 5500 XT is a good choice. The video card will be a good alternative to the outdated Radeon RX 580 and competes on equal terms with the Nvidia GeForce GTX 1650 Super, getting close even to the GTX 1660. In the budget segment, a lot depends on the price, and the AMD Radeon RX 5500 XT is ready to tear and throw, pushing the competition. GPU – Polaris 30 | GCN Cores – 2.304 | Memory – 8GB GDDR5 | Bus – 256-bit Pros : Cons : The latest generation of Polaris has given us 12nm chips with impressive gaming power at a very modest price tag. They are slightly inferior to the GTX Turing graphics cards, but still able to surprise. The latter has long been a problem for “red” GPUs, but you can easily find a card with additional cooling that will solve this misunderstanding. GPU – Polaris 20 | GCN Cores – 2.304 | Memory – 8GB GDDR5 | Bus – 256-bit Pros : The price of the RX 580 has gone down, and it has found itself in the budget segment. Its price tag may not be attractive enough for many, but the video card justifies every ruble invested in it. After all, it was AMD’s flagship just a couple of years ago. Actually, it is still capable of producing impressive results in 1080p and even 1440p. GPU – Polaris 20 | GCN Cores – 2.048 | Memory – 4GB GDDR5 | Bus – 256-bit Pros : Cons : An excellent budget graphics card that can make the GTX 1050 Ti shine with its presence. Still, the 4GB RX 570 can handle the highest settings at 1080p. A real budget hero. There are a number of criteria to consider when choosing a graphics card for your gaming PC. First of all – the price: in the current realities, the cost of a graphics adapter often has the greatest impact on decision making. When choosing between Nvidia and AMD, it’s worth considering that the “green” graphics cards offer flagship performance with ray tracing and DLSS intelligent anti-aliasing technology. AMD, in turn, is betting on a more attractive price, and its cards work better with AMD processors. Intel has already joined the race between video chip manufacturers, so the choice will expand soon. Based on your monitor resolution and desired graphics settings. When it comes to video memory, the “more is better” statement is true, especially when it comes to 4K gaming. Although there are nuances here too: a weak video chip combined with a large amount of memory will not give out impressive performance. In other words, you need a balance: do not be too lazy to study the characteristics and tests of the video card so as not to mistakenly buy an “office plug” with a large amount of memory that it does not need. Any graphics card consumes power, so it’s important to make sure your power supply matches the graphics adapter’s appetites. Usually, the specifications of the video card indicate the recommended power supply, and if it is weaker than required, it is better to take care of a replacement. Table of contents Toggle Best priced and features budget gaming cards in 2022 from Nvidia and AMD that can run the latest games. The popularity of mining has long gone and the cost of video cards is getting cheaper every day, which allows you to buy good cards at affordable prices. Nvidia’s Geforce GTX 1650 is an improved version of the previously popular graphics card GTX 1050 . Like the 1050, this model has a fairly low cost and very good features that allow you to play a large number of games with high FPS. GTX 1650 has 4GB VRAM , low power consumption in 75W , and frequency up to 1665MHz . On average, this video card will allow you to play in 1080p (Full HD) at 60-70 frames per second. But at 1440p, the card will allow you to run games with FPS in the region of 20-30 frames. The GeForce RTX 3050 is the cheapest graphics card from the manufacturer Nvidia , which supports ray tracing, as well as DLSS super resolution technology. This video card has very good characteristics for its low cost. 3050 has 8GB video memory , frequency up to 1777MHz , power consumption 135W , and one of the newest types of memory – GDDR6 . Thanks to these characteristics, this graphics card can provide stable 60+ FPS at high / ultra settings in such popular games as Fortnite (ultra 1080p about 60-70 FPS), Hogwarts Legacy (high 1080p 55-6 0 fps ), Atomic Heart (medium/high settings 60-75 FPS). AMD Radeon RX 6600 is a cut down version of RX 6600XT . It is also a competitor to the RTX 3060 graphics card from the Nvidia . This video card has 8 GB VRAM , frequency up to 2491 MHz and power consumption 132W . The card will allow you to play most games at high / ultra settings with an average of FPS 118 . Playing in 2K will get a maximum of 60 FPS . For example, in Cyberpunk 2077 the video card will allow you to play on ultra settings from 55-65 FPS . The GeForce RTX 3060 is one of the most popular models in the entire RTX line. It, like the RTX 3050 , supports ray tracing and DLSS technology, which will allow you to play any game at ultra settings with high FPS . This model has 12 GB video memory , frequency up to 1777 MHz and power consumption 170 W . RTX 3060 will run most modern games at high settings in 120+ FPS . For example, Battlefield 5 will run at high settings with a frame rate of 120-130 , Horizon Zero Dawn will run at the same high settings at 100-110 FPS . Fortnite on ultra settings will work in 120-130 FPS . This model allows you to play Full HD with high and stable FPS , as well as include additional graphic effects such as ray tracing. The video card has 8 GB video memory , frequency up to 2635 MHz and power consumption 176 W . Thanks to the powerful performance of this graphics card, you can run most modern games in stable 130 – 150 fps (1080p Full HD) .
The only problem is that its little sister RTX 2070 Super is almost as good in terms of price and power, so sometimes you might wonder why the RTX 2080 Super is needed? NVIDIA GeForce RTX 2070 SUPER
And considering the fact that the 2070 costs a couple of hundred dollars less than the 2080, you can imagine what a bargain this is. NVIDIA GeForce RTX 3060
In any case, this is a good “folk” map, which you will not have problems with most current AAA titles. NVIDIA GeForce RTX 2060 SUPER
NVIDIA GeForce RTX 3050
NVIDIA GeForce GTX 1660 TI
However, given the small number of games that really use ray tracing, this loss becomes not so significant. And it becomes even smaller if you remember at what cost for PC performance all this beauty is given. NVIDIA GeForce GTX 1660
For now. NVIDIA GeForce GTX 1650 SUPER
A consolation can be the impressive power efficiency of the graphics card, which gives room for creating quiet and economical gaming builds. NVIDIA GeForce GTX 1050 Ti
AMD
AMD Radeon RX 6900 XT
AMD Radeon RX 6800 XT
The video card is low-noise, supports the HDMI 2.1 standard, gives a performance boost when used with AMD Ryzen 5000 family processors – an excellent solution for enthusiasts, especially considering the democratic price. AMD Radeon RX 6700 XT
:max_bytes(150000):strip_icc()/how-to-use-xbox-game-pass-on-your-pc-featured-c6d74b87b3ea48bc9ec4c5cc6417aa63.jpg)
These are impressive figures, but only if you forget about ray tracing, where the “red” cards are traditionally in the laggard position, even though this model is equipped with hardware support for RT. AMD Radeon RX 6600 XT
It makes no sense to talk about something more serious, but the video card does not pretend to be: AMD did not declare that the RX 6600 XT would show itself worthy, for example, in 4K. AMD Radeon RX 5700 XT
Yes, the Nvidia card is more powerful, but the difference is incredibly small. AMD Radeon RX 6500 XT
AMD Radeon RX 5700
AMD Radeon RX 5600 XT
304 | Memory – 6GB GDDR6 | Bus – 256-bit
AMD Radeon RX 5500 XT
But only if you choose a model with 8 GB of video memory: alas, the 4 GB version does not perform in the best way even at this resolution. AMD Radeon RX 590
True, sometimes they can get hot… AMD Radeon RX 580
AMD Radeon RX 570
There is an 8GB version of the RX 570 on the market, however, the extra 4 GB mark-up brings it very close to the cost of the RX 580, which now, we must admit, looks very attractive. How to choose a gaming graphics card?
The higher the resolution, the more powerful graphics card you will need; the same is true for those who like to move all the graphic settings sliders to the right. In turn, if you plan to play undemanding network titles, you can think about a budget video card or a video chip (APU) built into the processor. 
Also pay attention to the dimensions of the video adapter, otherwise you may encounter difficulties when trying to install it in a PC case. Best budget gaming graphics cards in 2023
NVIDIA GeForce GTX 1650
Such characteristics are quite enough if you are not chasing the maximum stable FPS and a beautiful picture. NVIDIA GeForce RTX 3050
AMD Radeon RX 6600
NVIDIA GeForce RTX 3060
AMD Radeon RX 6650XT
