- Graphics cards are mini-computers you install into your computer to handle the graphical processing. They’re designed to run as an add-on to your computer build.
- The memory bus is the amount of data that can be transferred and processed by the card in one instance. The wider the memory bus, the more data can be processed.
- The clock speed of a processing unit is how many times a processor can complete a data processing cycle per second. It’s measured in Giga-Hertz (GHz) nowadays.
While the early 2020s signaled a massive crash in the graphics card market, the market has largely recovered, and high-end graphics cards are now freely available once again. With that comes a surge of interest in the graphics card industry and what they produce. Thus, History-Computer has compared and ranked the best GPUs in 2023 for your research and viewing pleasure. The list may surprise you!
There are many different methods we use to determine the best GPUs in 2023. To understand what makes a high-end graphics card the best of the best, you should first understand the inner mechanisms that make up a graphics card. Graphics cards are mini-computers you install into your computer to handle the graphical processing.
Since the graphics processor is strictly about processing graphics, it runs as an add-on to your computer build. However, some workarounds can process logic using your graphics card, such as those seen in GPU cryptocurrency mining.
Here are the most important metrics you want to look at when choosing a graphics card.
Width of the Memory Bus
One way to measure the speed and relative power of your graphics card is by looking at the size of the cards measured in bits. The memory bus is the amount of data that can be transferred and processed by the card in one instance. The wider the memory bus, the more data can be it can process.
Making sure your computer’s architecture is compatible with your graphics card’s memory bus is essential. 128-bit and 256-bit memory buses used to be the standard, and they are the memory buses with which most motherboards are compatible. However, larger memory buses, up to 4096-bits, are becoming more popular and common amongst graphics card manufacturers trying to push the envelope technologically.
Typically modern graphics cards have larger memory buses. So, you’ll want to ensure your architecture is up-to-date to avoid the computer’s interfaces throttling the card.
Video RAM Size
Graphics cards come loaded with dedicated RAM used solely for processing graphical information. Video RAM, or VRAM, has individual metrics to determine how strong and fast it is. In addition, the amount of VRAM determines how much space your graphics card has to work with when processing multiple graphical states. This can be extremely important for gamers who will be processing massive amounts of graphics and visual conditions to form the game’s different levels, screens, and conditions.
The more VRAM your card comes with, the more data about the graphics it will be able to store and recall while you game. Therefore, VRAM size is essential for competitive game players who need their computers to store and recall real-time game data as they play.
The clock speed of a processing unit is how many times a processor can complete a data processing cycle per second. We measure this in Giga-Hertz (GHz) nowadays. So 1 GHz represents 1 billion processing cycles per second… that’s a lot of data moving around!
Typically, the higher the clock speed, the faster the unit can process data. In the case of graphics cards, the data being processed is the information and states of the graphical information being displayed by programs on your computer.
It’s also worth mentioning that clock speed is not a good indicator of a processor’s practical speed. While the clock speed is necessary, the practical speed of a processor is a combination of the clock speed and its interfacing speed with other components. So, while getting a unit with a high clock speed is important, it’s also vital to ensure that your computer’s interfacing won’t artificially slow down the processor.
Another factor you’ll want to consider when choosing a graphics card is what outputs are available to the card. The outputs are how the card interfaces with the monitors to display the data it processes.
Older graphics cards will be outfitted with depreciated outputs, making them incompatible with some newer monitors. Conversely, older monitors may also be incompatible with more recent graphics cards. So, you’ll want to be sure you’ve selected a graphics card compatible with your monitors.
Our Best Choices Condensed
- Best Overall: Nvidia RTX 4080
- Best Budget Option: Nvidia GTX 1660 Super
- Best Midrange Option: Nvidia RTX 3060ti 12GB
- Best AMD Option: Radeon RX 6800 XT
- Best Mid-Range Card: NVIDIA RTX 3060
- Best Premium: NVIDIA RTX 4090
- Best Budget AMD Card: Radeon RX 6750 XT
- Best Intel Graphics Card: Arc A770 Limited Edition
Best Overall: NVIDIA RTX 4080
- 4th Gen Tensor Cores: 2x AI performance.
- 3rd Gen RT Cores: 2x ray tracing performance.
- Boost Clock: 2520 MHz, 16GB GDDR6X, 256-bit, 22.4 Gbps.
If you’re looking for a good GPU but don’t want to shell out for the RTX 4090, then the RTX 4080 from Zotac is an excellent upgrade to your PC without breaking the bank, and that’s why we’ve chosen it as one of our best GPUs in 2023.
The 4080 features NVIDIA’s latest generation Tensor cores for even better AI processing and upscaling. It also features 3rd generation ray tracing that uses 76 ray tracing cores, 304 Tensor cores, 16 GB of VRAM, and a 256-bit memory bus width.
What makes the Zotac version of this card special is the lower price tag than the Founder’s edition. This means you get all the performance benefits coupled with the added benefit of 3rd-party cooling and PCB design.
Check out Nvidia RTX 4080 on Amazon.
Best Budget Option: Nvidia GTX 1660 Super
- 6 GB Graphics RAM
- Super compACt 6.83-inch card
- 3 x DisplayPort 1.4, 1 x HDMI 2.0b
- Boost Clock: 1785 MHz
The Nvidia GTX 1660 Super emerges as an unrivaled contender in the mid-range GPU market segment. This compact powerhouse is built on the new Turing architecture. It features 6GB 192-bit GDDR6 memory that easily outpaces GPUs in similar price brackets.
Despite its impressively compact size of just 6.83 inches – which fits 99% of systems – this dual-slot GPU doesn’t compromise on performance. It is capable of delivering 4K, HDR, and VR-ready capabilities.
When it comes to real-world application, the GTX 1660 shines. Its adeptness at handling graphically intensive games like Roblox is commendable. Not only does it swiftly upload graphics, but it also significantly enhances the depth and brightness of colors, offering an immersive gaming experience.
Additionally, its gaming prowess isn’t limited to casual gaming; it produces surprisingly high frames even in competitive games like Fortnite when paired with a decent CPU and RAM.
Beyond gaming, the GTX 1660 also stands out for its ability to expedite workflows, particularly in the realm of photo editing. It can drastically reduce rendering time, for example, by approximately 40% when used with software like Topaz Sharpen, providing a cost-effective solution for streamlining creative processes. All these factors make the Nvidia GTX 1660 an excellent value for money and the best midrange GPU on the market.
Check out the GTX 1660 Super on Amazon.
Best Midrange Option: Nvidia RTX 3060ti 12GB
- NVIDIA Ampere architecture with ray tracing and tensor cores
- 12GB GDDR6 VRAM, PCIe 4.0 support, 1807 MHz boost clock
- IceStorm 2.0 Cooling, active fan control, metal backplate
- DisplayPort 1.4a, HDMI 2.1, 8K resolution, DirectX 12 Ultimate support
The NVIDIA GeForce RTX 3060 Ti is a notable offering in the GPU market, characterized by an upgrade to 12GB of VRAM as compared to the usual 8GB found in most RTX 3060 models.
This card harnesses the power of NVIDIA’s Ampere architecture, with 2nd generation Ray Tracing Cores and 3rd generation Tensor Cores, providing a marked boost in graphical computing performance. Its 12GB 192-bit GDDR6 memory is capable of reaching 15 Gbps, while its GPU clock speed peaks at 1807 MHz.
The Zotac edition sports tasteful upgrades such as IceStorm 2.0 Cooling, Active Fan Control, and Freeze Fan Stop technologies, all housed under a durable metal backplate.
Many users reported a significant improvement when upgrading from previous models, such as the 1660 Super, to the RTX 3060 Ti. The boost in VRAM from 6GB to 12GB not only dramatically enhances gaming performance, but also leads to better benchmark scores across the board.
While it is a great budget offering, the superior performance of the RTX 3060 Ti can be underutilized due if the rest of your PC isn’t properly specced out or configured. Don’t underestimate this card because of the low price tag: it is a gaming beast.
Check out the Nvidia RTX 3060ti 12GB on Amazon.
Best AMD Option: Radeon RX 6800 XT
- Memory Speed:16.0 Gbps.Digital Max Resolution:7680×4320
- AMD Radeon 6800XT
- Video Memory: 16GB GDDR6
- Stream Processor: 4608.Avoid using unofficial software
- Game Clock: 2065MHz (OC) / 2015MHz (Silent)
AMD is a household name at this point, at least to PC enthusiasts. They are an excellent budget component manufacturer, and they made several of the best GPUs in 2023.
Plus, it’s hard to beat their processors when you don’t have a lot of money to throw around. The Radeon RX 6800 XT is an excellent high-end card by AMD.
The 6800 XT is a bit beefier than the RTX 3080, but your mileage will vary based on your other components and how they interface with the card. The best part about this card is its price-to-performance ratio. In other words, you get a lot of bang for your buck.
One major downside of the 6800 XT is its size. The 6800 XT is a 2.5-slot card, meaning if your computer is on the smaller side, you may need to upgrade your case to use this card. In addition, smaller builds won’t have the space the card needs to sit during installation, let alone use.
Check out Radeon RX 6800 XT on Amazon.
Best Premium: NVIDIA RTX 4090
- Chipset: GeForce RTX 4090
- Video Memory: 24GB GDDR6X
- Memory Interface: 384-bit
- Output: DisplayPort x 3 (v1.4a) / HDMI 2.1 x 1
- Digital maximum resolution: 7680 x 4320
Despite several months since its initial release in October, the NVIDIA RTX 4090 continues to set the gold standard in the world of high-performance graphics processing.
This absolute unit is outfitted with an impressive 16,384 CUDA cores, making it one of the most powerful GPUs on the market, unmatched in its ability to handle intensive graphical workloads. It supports 4K 120Hz HDR and 8K 60Hz HDR output, ensuring vibrant, lifelike visuals for even the most demanding games and creative applications.
The 4090 uses NVIDIA’s new Ada Lovelace architecture to bring the most excellent graphics to your home PC. In addition, it features dedicated ray tracing cores for gamers addicted to ray traced games (no hate, they’re gorgeous and well worth being addicted to!)
The card also features AI-accelerated graphics, allowing the processor to process graphics faster than ever. This feature takes full advantage of NVIDIA’s low-latency Reflex architecture to bring you the smoothest performance no matter your game.
Check out NVIDIA RTX 4090 on Amazon.
Best Budget AMD Card: Radeon RX 6750 XT
- Chipset: AMD RX 6750
- Memory: 12GB GDDR6
- Cooling: XFX QICK 319 Triple Fan
- Boost Clock: Up To 2623MHz
If you’re looking for an AMD card but don’t have the space for the 2.5-slot 6800 XT, the Radeon RX 6750 XT is an excellent choice. Unfortunately, the 6750 XT is a step down from the 6800 XT. Still, it offers a significant power boost to many older PCs without breaking the bank or upgrading your case to accommodate a massive card.
The RTX 6750 XT is a mid-high graphics card. So, you’ll be able to play many of the newest games on high graphics settings. But you may experience a bit of a cap with the highest possible settings, as the 6750 XT doesn’t have the same power as the 6800 XT.
Check out Radeon RX 6750 XT on Amazon.
Best Intel Graphics Card: ARC 770 Limited Edition
- 16 GB of GDDR6
- HDMI 2.1 with VRR
- Xe HPG Architecture
- Real-time ray tracing
- PCI Express 4.0 (x16)
- 512 GB/s
They are the new kid on the block, at least as far as graphics cards go. While Intel has become a household name in the processor market, their GPU offerings never amounted to much until recently.
With the launch of the Intel Arc A770, all of that is about to change. AMD and Nvidia finally have some competition from a noteworthy contender.
The Intel Arc A770 is a high-performance graphics card leveraging Intel’s Alchemist architecture. The GPU comes with 32 Xe cores, and 512 XMX engines across 8 render slices, which together offer formidable computing power for gaming and other graphically intensive tasks.
The A770 also boasts 32 ray tracing units for realistic lighting and reflections, supporting the latest advancements in graphics technology.
With a graphics clock speed of 2100 MHz and 16GB of GDDR6 memory running at 17.5 Gbps across a 256-bit interface, this GPU provides a massive memory bandwidth of 560 GB/s, ensuring seamless performance even in demanding scenarios. It uses the PCIe Gen 4.0 x 16 system interface and has a TDP of 225W, so you’ll easily slot it into your current system.
Check out Intel Arc A770 on Amazon.
How to Choose the Best Graphics Card
The best way to choose a graphics card is to look at benchmarks. Typically, manufacturers will perform a series of benchmarks that they can display. Still, it’s important to remember that these are not the same as practical user benchmarks. Benchmarks are an excellent way to determine which card is the best GPU in 2023.
Benchmarking a graphics card is related to the graphics card itself and the other components in your computer and how they interface and interact with your graphics card. Merely having the newest graphics card alone won’t guarantee the highest fidelity graphics and performance. Computers are going to be limited by their oldest parts.
This is called a “bottleneck” since the lowest performing component will hold up the rest of your PC from running its best.
So, if your motherboard and CPU are ancient, your build is going to struggle – no matter how nice your GPU is.
This variance is why looking at user benchmarks is crucial when shopping for your GPU. Manufacturer benchmarks are done in the best possible conditions with brand-new high-end workstations to give the company the ideal numbers to display on their ads. However, practical user benchmarks provide a more realistic idea of how a graphics card will perform under load.
You can benchmark your GPU using specific benchmarking software or playing a game and recording your FPS. Before purchasing a graphics card, we recommend checking several benchmark aggregators like UserBenchmark, and Benchmark.UL, and Tom’s Hardware. These benchmarks will give you an idea of how your prospective purchase will function and benchmark when used in a non-ideal build.
On your own PC, you can use programs like FRAPS to track your framerate in-game and track resource usage to see whether the games you’re playing use more resources than they usually do.
You’ll also want to choose your new graphics card based on the size of your computer case. As graphics cards get more powerful, they also become larger and more unwieldy. Computers with smaller issues cannot fit newer graphics cards into their measurements.
People with mini, micro, and mid-sized towers must be cautious when choosing new parts, as these cases tend to be too small for some of the newer internal PC components. Cards like the 6800 XT require more than the typical two PCIe slots. So if your case doesn’t have space for a two-slot card, it certainly won’t fit a card larger than that.
It’s also important to remember that your GPU’s output panel needs to extend out through the back of your case. Even if there’s enough raw space inside the case, there may not be enough panel outputs on the back of the case to accommodate the GPU.
You’ll also want to choose a GPU that is internally compatible with your other components. While newer components tend to be backwards compatible, a GPU plugged into a motherboard with a smaller memory bus than the GPU is capable of running will be artificially bottlenecked by the motherboard.
Additionally, you’ll need to ensure that your power supply has a high enough rating to be compatible with your GPU. Newer GPUs have immense power draws, and power supplies can and will break and become entirely defunct when presented with components that draw too much power.
What to Know Before Buying a GPU
There are several major considerations you should make before upgrading your GPU. The first major consideration you should look at is whether your current hardware is capable of getting the most out of your new GPU. If you have a lot of older hardware in your computer, the hardware may artificially bottleneck your GPU.
For instance, if your PCIe slots only have a memory bus of 256-bits, you won’t be able to take advantage of the expanded memory bus that comes with newer graphics cards. The best GPUs in 2023 typically have beefier specifications that are compatible with older hardware. So if your hardware is too old to take advantage of a new graphics card, you might want to hold off until you can upgrade more of your hardware with it.
You’ll also want to make sure your peripherals are powerful enough to display the upgraded graphics information. Monitors are only able to refresh the picture–draw a new frame–a limited number of times per second.
For instance, a 60 Hertz monitor can redraw frames up to 60 times per second. But if you’re trying to run games at a higher framerate than that, a 60 Hz monitor can’t keep up, and you won’t see the additional frames and smoothness.
What It’s Like to Use a GPU
The truth is that unless you’re doing graphically intensive tasks like graphic design, digital art, video editing, or gaming, you probably won’t notice a difference between having a GPU and not having one. But if you weren’t doing something graphically intensive, you probably wouldn’t be here in the first place.
Using a graphics card when doing graphically intensive tasks will typically result in smoother performance, less load on your CPU, and buttery smooth graphics. Those performing graphically intensive tasks should invest in a graphics card if it’s possible to upgrade your graphics chipset (Mac users will not be able to do this at all.)
Using one of the best GPUs in 2023 will give you an experience you’ve never had before if you’re used to using older hardware. The graphics will be smooth, detailed, and beautiful like you’ve never seen them before.
An updated graphics chipset is an excellent choice for improving your computer’s performance under heavy loads. Upgrading your chipset can help bolster your computer’s ability to process graphics without straining your other components. Finding the right graphics card for your needs is essential when upgrading your computer. So, do much research on the model you plan to buy and its useful benchmarks to ensure that your computer will run the way you expect it to!
- How Much of Tesla Does Elon Musk Actually Own? As Elon Musk sells off more and more of his shares in Tesla, the question bears revisiting.
- How to Edit Videos on Windows 11 (with Photos) Got a video you need to edit? Here’s a guide to do your editing on Windows 11.
- Who Is Responsible When Self-Driving EVs Get Into Accidents? As self-driving cars become more common, accidents will become a liability issue. Who would be at fault?