10 best graphics cards for gaming

Review of the best according to the editorial board. On the selection criteria. This material is subjective and does not constitute advertising and does not serve as a purchase guide. Before buying, you need to consult with a specialist.

In a gaming computer, the graphics card matters even more than the processor, within reason. And for the most demanding gamers, the cost of a suitable model can make up the lion's share of the entire gaming station assembly, even including the monitor. But for the mass of less sophisticated users, significantly more modest characteristics will be enough.

The editorial staff and analysts of the RankQuality service have prepared for you an overview of the most relevant and demanded gaming video cards at the beginning of 2020, covering all reasonable price categories. Simple guidelines will help you equip your computer with a card sufficient for your entertainment needs without overpaying for unnecessary functionality.

Rating of the best video cards for gaming

Nomination	a place	Name of product	price
The best budget graphics cards for gaming	1	MSI Radeon RX 560 1196MHz 4096MB 7000MHz	RUB 7,516
	2	GIGABYTE GeForce GTX 1050 Ti 1316MHz 4096MB 7008MHz	RUB 9,300
	3	GIGABYTE GeForce GTX 1660 Ti 1860MHz 6144MB 12000MHz	RUB 20,180
The best mid-range graphics cards for gaming	1	GIGABYTE Radeon RX 5700 1565MHz PCI-E 4.0 8192MB 14000MHz	RUB 24,459
	2	MSI GeForce GTX 1070 Ti 1607MHz 8192MB 8008MHz	RUB 29,990
	3	MSI GeForce RTX 2060 SUPER 1695MHz 8192MB 14000MHz	RUB 29 375
	4	ASUS ROG GeForce RTX 2070 1410MHz 8192MB 14000MHz 256 bit Strix Gaming OC	RUB 36 330
The best top graphics cards for gaming	1	Palit GeForce RTX 2070 SUPER 1605MHz 8192MB 14000MHz	36 900
	2	MSI GeForce RTX 2080 SUPER 1845MHz 8192MB 15500MHz GAMING X TRIO	RUB 51,665
	3	GIGABYTE GeForce RTX 2080 SUPER 1860MHz 8192MB 15500MHz AORUS	RUB 52,500

The best budget graphics cards for gaming

The first collection of the review presents the most inexpensive and, accordingly, the least productive video cards that are able to somehow “pull” modern video games. These are models announced at least 2-3 years ago. Hereinafter, in terms of 'pure' technical characteristics, similar video cards are in the assortment of each of the top manufacturers (MSI, GIGABYTE, ASUS, Palit) and they are essentially identical. And our experts made the choice of this or that manufacturer based on feedback from real users. The degree of 'budget' in this particular review is limited to the upper price limit of 20 thousand rubles.

MSI Radeon RX 560 1196MHz 4096MB 7000MHz

Rating: 4.7

And it opens an overview of the smallest possible solution for gaming in 2020. Of course, no one forbids taking something weaker, but then you can definitely forget about a comfortable game. Radeon RX 560 is that conditional minimum that will at least give tolerable fps in the most popular games at at least FullHD resolution.

The video card is equipped with a Polaris 21 graphics processor with a frequency of 1196MHz. In the context of this particular model, this point is especially important, as it speaks of a real RX 560 with support for 64 texture units and 1024 stream processors. Outwardly almost indistinguishable, the RX 560D is the de facto previous generation – the RX 460 with the Baffin Pro chip.

RAM onboard – 4GB GDDR 5th generation with an effective frequency of 7GHz and a bus width of 128bit. SLI / CrossFire supported. Characteristics of the processor mathematics: 5th version of shaders; already mentioned 1024 arithmetic logic units (ALU); 64 texture units; 16 rasterization modules; anisotropic filtering – up to 16x maximum. The limit of possible graphics resolution is 7680×4320. Supports up to three monitors simultaneously.

The device is relatively compact – 155x113x39mm, but it still takes up two slots. The chip is cooled by a relatively large fan in one copy, due to which the card is also quite quiet. Heat dissipation is estimated at 80W. For normal operation, it is recommended to supply power from a power supply unit with a power of at least 400W, and better – at least 450W.

From the interface side, the characteristics are as follows: DisplayPort version 1.4, DVI-D, HDMI 2.0b, HDCP support. In the firmware part, we see support for OpenGL4.5, DirectX12, Vulkan, OpenCL2.0, AMD APP (ATI Stream). The card itself connects to a PCI-E 16x 3.0 motherboard slot.

Independent test results show that at FullHD 1920 × 1080 resolution, the card pulls fps above 60 in games such as Call of Duty MW (2019), Battlefield V (2018), Doom (2016), DiRT Rally (2015), Final Fantasy XV ( 2018). This is, of course, provided that the system has a sufficiently powerful processor, since things will be bad on some Celeron.

Advantages

• minimum heating;
• quiet work;
• compact size;
• reliability and reliability;
• no additional food is needed;
• the lowest possible price for a video card that is more or less adequate to modern requirements.

disadvantages

• fan rattling – it is treated with a slight modification and sealing;
• It is better to replace the regular thermal grease on the chip with a more productive one.

GIGABYTE GeForce GTX 1050 Ti 1316MHz 4096MB 7008MHz

Rating: 4.6

The second issue is a slightly more expensive and slightly more productive video card based on the GeForce GTX 1050 chip. The manufacturer this time is GIGABYTE.

This model does not have any record advantages over the previous one, and in general its estimated performance is approximately 30% higher. For a significant part of users, this difference does not cost one and a half to two times the price, but it can be significant if it significantly increases comfort in any particular favorite game.

This model is based on the legendary NVIDIA Pascal architecture, which is considered to be the most advanced in the world for its time. The chip itself is created according to the most advanced 14nm process technology widely used in 2018-2019. However, this also applies to the previous model.

The processor frequency in this model is 1316MHz. In Boost mode, which is most often automatically accompanied by Gaming Mode, the frequency reaches 1430 MHz. In the operating system mode, the card can also be overclocked by standard means, and even more – up to 1455 MHz. Equipped with 4GB GDDR5 RAM with a frequency of 7008MHz and a 128-bit bus. The maximum resolution is 7680×4320.

Mathematical block parameters: 768 universal processors; 48 texture units; 32 rasterization blocks; 5th version of shaders; 16x anisotropic filtering. Software and firmware parameters: support for OpenGL4.5, DirectX12, CUDA version 6.1, Vulkan, OpenCL1.2.

The interface configuration of this model is almost exactly the same as the previous one: DisplayPort 1.4, HDMI2.0b, DVI-D, HDCP support. Supports three monitors in simultaneous operation mode. The minimum recommended power supply in the system unit is 300W. The dimensions of the device are 191x111x36mm, it takes up two slots. Two fans are used for cooling.

Compared to the previous model, this graphics card runs louder and gets hotter. But all this is within reason – the noise is almost not felt with rare exceptions, the heating is only formally high and does not cause throttling, if, of course, everything is in order with the fans and thermal grease.

Advantages

• Quiet enough work;
• advanced NVIDIA Pascal architecture at the core;
• price-performance ratio;
• support for three monitors;
• turbo mode;
• low requirements for the power supply.

disadvantages

• a small but noticeable percentage of complaints about the quality of materials and workmanship.

GIGABYTE GeForce GTX 1660 Ti 1860MHz 6144MB 12000MHz

Rating: 4.6

The third and final position in the magazine's selection of the best budget graphics cards for gaming is another product from the GIGABYTE brand with the “heart” of nVidia – the GeForce GTX 1660 Ti chip.

First, let's clarify what the Ti index in the name means. Here, this designation does not mean a stripped-down version of the GeForce GTX 1660, as some mistakenly believe, but on the contrary – a more powerful one. Only this power is achieved not by direct factory overclocking, but by increasing the number of execution units, that is, CUDA cores, as the developers themselves call them.

The first noticeable advantage over the previous model, in addition to the generation itself, is the increased amount of RAM from 4 to 6 GB. Further study of the characteristics firmly convinces that we are dealing with a very decent device. So, the TU116-400-A1 graphic processor, created according to the 12nm process technology, operates at a frequency from 1.5 GHz to a maximum of 1.965 GHz – within these limits, overclocking is allowed by standard means. The memory bus is 192-bit wide. The memory here, by the way, is very advanced – GDDR6 with an effective frequency of 12 GHz (physical – 3GHz).

The total number of arithmetic logic units in the processor is 1536. There are 96 texture units and 48 rasterization units. The shaders version is 6.1. Maximum Resolution: 7680 × 4320 / 30Hz via DisplayPort; 3840 × 2160 / 60Hz via HDMI; 2560 × 1600 / 60Hz or 1920 × 1200 / 120Hz via Dual-Link DVI; 1920 × 1200 / 60Hz or 1280 × 1024 / 85Hz via Single-Link DVI. The available outputs are one HDMI 2.0b and three DisplayPort1.4 outputs. Supports simultaneous work on four monitors. On the software side, we see support for OpenGL version 4.6, DirectX 12th generation, CUDA version 7.5, Vulkan, OpenCL version 1.2.

This video card is already quite large – 280x117x40mm, and requires additional power supply via an 8-pin connector. The power requirement for the power supply in the gaming PC system unit is at least 450W. The cooling system – three fans, more than copes with its task.

With the proprietary Aorus Engine utility, you can easily control the operating modes of the video card, including the degree of overclocking of the GPU.

Advantages

• performance in terms of cost;
• quiet work;
• low heating with a large margin to the maximum;
• 6GB of RAM;
• technical process 12nm (processor only);
• GDDR6 memory;
• beautiful lighting.

disadvantages

• many users are 'nostalgic' for the DVI output, which this model lacks.

The best mid-range graphics cards for gaming

The second selection in the review of the best video cards for games according to the magazine's version is more expanded and represents the most massive segment – gaming video cards of the middle price category up to 30 thousand rubles at the beginning of 2020. Almost any of them is capable of more than covering all the leisure needs of an ordinary user and most of them for an advanced gamer.

GIGABYTE Radeon RX 5700 1565MHz PCI-E 4.0 8192MB 14000MHz

Rating: 4.9

The selection is opened by a video card made by GIGABYTE on the Radeon RX 5700 platform. This is the only model in our review that connects to the PCI-E 4.0 interface of the motherboard, and not 3.0, like everyone else.

In the conditional middle class, this is one of the most productive video cards, and it is the cheapest among the models that follow in the list. The 'heart' of the device is a powerful AMD Navi 10 processor, created according to the most advanced 7nm technical process. The processor frequency is 1656MHz, and there is still a small potential for 'legal' overclocking. In general, the RDNA microarchitecture used here provides a large margin for comfortable playing modern video games. At a FullHD resolution of 1920×1080, it provides a high frame rate at maximum graphics settings. In some, it doesn't even crash at 2560×1440.

Let's describe the technical characteristics of the device in its purest form. The volume of video memory GDDR6 here is 8GB. The effective memory frequency is 14GHz. Bus width – 256bit. SLI / CrossFire supported. Mathematical block parameters: 2304 arithmetic-logic devices; 144 texture units; 64 blocks of rasterization.

The maximum resolution that the card can provide is 7680×4320. Simultaneous work on 4 monitors is supported. For this, there are HDMI 2.0b interfaces and three DisplayPort 1.4 ports. HDCP supported. On the software side, there is support for OpenGL4.6, DirectX12, OpenCL2.0, Vulkan.

The dimensions of the video card are quite impressive – 300x114x50mm, it takes up three slots with its volume and needs additional power supply via an 8-pin connector. The recommended power of the power supply in the system unit is 600W.

The three-slot cooling system WindForce 3X deserves special praise in this model. Three fans are quiet enough to effectively keep the CPU temperature within safe limits. And the processor itself does not heat up too intensely due to the same 7nm process technology. But in 4K resolution, there is already a noticeable noise.

Advantages

• 7nm process technology;
• sufficiently high performance;
• relatively quiet (up to 4K);
• GDDR6 memory with you;
• efficient cooling system WindForce 3X;
• not too flashy, but nice lighting.

disadvantages

• dimensions – occupies three slots;
• when playing in 4K resolution, it makes a noticeable noise.

MSI GeForce GTX 1070 Ti 1607MHz 8192MB 8008MHz

Rating: 4.8

The review continues with a video card manufactured by MSI based on GeForce GTX 1070 Ti. What the Ti index means, we have already clarified above – increased power compared to similar models without this index.

We also emphasize that this model is not very new, but it is still in demand and costs decent money. There are a number of reasons for this: relative compactness, predictability, the test of time and a number of other points that are important for certain categories of users (for example, the presence of a full-fledged RGB backlight).

Let's move on to the technical specifications. The GP104-300-A1 graphics processor with a nominal frequency of 1607MHz was created using the 16nm process technology. This is obviously outdated technology, especially against the background of 7nm in the previous example, but for most consumers this is not a very tough barrier. The same can be said about the memory – here it is GDDR5 at 8008MHz, but the volume is acceptable – 8GB. Bus width – 256bit. SLI / CrossFire supported.

Mathematical block parameters: 2432 ALU; 5th version of shaders; 152 texture units; 64 rasterization blocks; 16x (maximum) anisotropic filtering. Software support: OpenGL4.6, DirectX12, CUDA version 6.1, Vulkan, OpenCL1.2. The maximum resolution supported by the card is 7680×4320. Up to four monitors can be connected at the same time. Interfaces: HDMI2.0b, DVI-D port, three DisplayPort version 1.4 ports.

Card dimensions – 279x140x42mm. Visually, it is not too smaller than the previous one, but in fact it takes up only two slots, not three. Requires additional 180W power via 8-pin connector. The recommended power of the computer power supply is 500W. Cooling is provided by a tandem of two efficient and not too loud fans.

Advantages

• high performance;
• effective cooling;
• relatively compact size;
• time-tested, predictable and reliable;
• RGB backlighting;
• there is a DVI-D port.

disadvantages

• obsolescence of individual elements and characteristics.

MSI GeForce RTX 2060 SUPER 1695MHz 8192MB 14000MHz

Rating: 4.7

The third number in the selection of the best mid-range gaming video cards according to the version, consider another interesting model from MSI. The company's assortment includes a cheaper modification without the SUPER prefix, but in this case this is not marketing, and it certainly makes sense to pay extra for the SUPER version from above.

In terms of performance, this model already allows you to practically not think about the graphics settings in modern video games, at least in FullHD resolution. The TU106-410-A1 graphics processor is based on the 12nm process technology, operates at a nominal frequency of 1695MHz with the possibility of slight overclocking. System Memory – 8GB GDDR6 with 14GHz effective frequency. The bus width is 256 bits.

The maximum resolution here is the same as in all previous models – up to 7680×4320. Simultaneous work on 4 monitors is supported. Standard interfaces: HDMI2.0b, three DisplayPort1.4 ports, HDCP supported. From the firmware side: support for OpenGL version 4.6, DirectX12, OpenCL 1.2, CUDA new version 7.5. Mathematical block parameters: 2176 ALU, new version of shaders – 6.1, 136 texture units, 64 rasterization units.

The dimensions of the device are 248x128x52mm, and it occupies three motherboard slots. Needs additional power via 8-pin connector and draws up to 175W. The recommended power of the computer power supply is 550W.

In general, this is a very decent solution for gaming, but it is not without some disadvantages. Not too critical, though. So, if this model is shorter in length than the previous one, then the thickness already makes you 'forget' about the three slots of the motherboard. Also, many users experience problems installing official drivers, especially on systems of the family Windows below version 10. Another thing is that at the beginning of 2020 the company Microsoft has already officially stopped support Windows 7, but the user base of this version remains huge. And one more thing – the card heats up noticeably even in the background. Heating is not dangerous, but it causes discomfort for some users.

Advantages

• high performance;
• quiet enough;
• effective cooling;
• support for up to 4 monitors simultaneously;
• reliability and stability in work (after solving possible problems with drivers);
• stylish design.

disadvantages

• there are problems with drivers;
• takes up three slots;
• heats up noticeably in the background.

ASUS ROG GeForce RTX 2070 1410MHz 8192MB 14000MHz 256 bit Strix Gaming OC

Rating: 4.7

And the first in our review is the representative of the bright products of the famous ASUS trademark. This is a 'pre-top' model based on the GeForce RTX 2070. It is part of the famous ROG (Republic of Gamers) series.

In 2019, the GeForce RTX 2070 series of graphics platforms received a serious and tangible update, and the improved versions received the Super prefix in the name. The changes were really noticeable, and this caused even a certain stir in the gaming environment. But that hasn't overshadowed the 'regular' RTX 2070s, and they are in the same demand. The closest example is the video card in question.

Key technical characteristics: graphic processor TU106-400A-A1, created according to the 12nm technical process with a nominal frequency of 1410MHz; video memory – 8GB GDDR6 with an effective frequency of 14GHz; bus width – 256 bits. In the math block we find 2304 ALU, 6.1 version of shaders; 144 texture blocks 64 rasterization modules. Firmware side: OpenGL generation 4.6; DirectX version 12; support for CUDA new version 7.1; OpenCL 1.2; Vulkan.

The maximum resolution provided by the card is 7680×4320. Supports four monitors in simultaneous operation. Standard interfaces: two ports HDMI 2.0b; two ports DisplayPort version 1.4; USB Type-C port.

The dimensions of the card are impressive – 305x131x49mm, three fans work for cooling, and the device takes up space as expected – three slots. Additional power is required, for this there are two connectors at once – 8-pin and 6-pin. The recommended power of the computer power supply is at least 550W.

This model, being a representative of the large ROG Strix series, has all the characteristic advantages of its neighbors in the lineup: improved PCB, factory overclocking, highly efficient cooling.

Advantages

• high performance;
• significant overclocking potential;
• reliability and reliability;
• efficient and flexible controlled cooling;
• USB Type-C;
• beautiful lighting.

disadvantages

• audible squeak of chokes and howling of fans at high speeds;
• possible problems with the drivers.

The best top graphics cards for gaming

Now we turn to the most interesting part of our review, where we will consider a selection of the best video cards of the top segment according to our experts. Among the top three models, we will analyze the newest solutions at a price of up to 60 thousand rubles. We will not consider the highest class due to the clearly inflated prices (80-90 thousand rubles) with not too impressive advantages for such a difference in price.

Palit GeForce RTX 2070 SUPER 1605MHz 8192MB 14000MHz

Rating: 4.9

First, we will consider the very affordable top-end video card of the Palit trademark. In terms of price, this model balances on the edge of the middle and top segment, but in terms of characteristics it definitely belongs to the conditional gaming Olympus.

Due to the fact that the manufacturer in this case is not as 'pretentious' as the well-known market leaders, he withstands the competition by minimizing the price. That is why this model shows in the top segment almost the best ratio of performance and cost as of the beginning of 2020.

Let's take a closer look at the technical specifications. The graphics processor – TU104-410-A1, created according to the 12nm process technology, operates at a nominal frequency of 1605MHz with overclocking capability. Video Memory – 8GB GDDR of the sixth generation with an effective frequency of 14 GHz and a 256-bit bus. There is built-in support for NVLink technology (ultra-efficient high-speed data transfer). The mathematical part has the following characteristics: 2560 ALU, shaders version 6.1; 160 texture units; 64 blocks of rasterization. Support for software standards: OpenGL version 4.6; DirectX12; support for CUDA new generation 7.5; Vulkan and OpenCL1.2 support.

The maximum possible resolution for this model is the same as for all others in the review – 7680×4320. Supports simultaneous connection and broadcast to four monitors. Interfaces – HDMI 2.0b and three DisplayPort 1.4 ports. HDCP supported.

Dimensions of the video card – 292x130x60mm. Cooling is provided by a pair of efficient and not too loud fans. In terms of volume, the device occupies three slots on the motherboard. To connect additional power, there are two 8-pin connectors. The recommended power supply unit is 650W.

This video card is large and heavy, many motherboards are deformed and there is a non-zero risk of motherboard failure. To prevent this from happening, it is enough to immediately pay attention to the possible deflection and take simple measures. Still, some users do not like the gluttony of the video card in terms of power consumption.

The official manufacturer's warranty for this model is 3 years.

Advantages

• ratio of price, capabilities and technological level;
• support for NVLink technology;
• overclocking potential;
• effective cooling;
• colorful lighting;
• quiet work;
• 3 years warranty.

disadvantages

• requires a lot of nutrition;
• heavy weight – bends the motherboard.

MSI GeForce RTX 2080 SUPER 1845MHz 8192MB 15500MHz GAMING X TRIO

Rating: 4.9

The second number in this collection is to draw your attention to an even more advanced solution with a chip from nVidia, this time from MSI. This is a powerful and very expensive graphics card based on the GeForce RTX 2080 SUPER platform.

The RTX 2080 Super modification is a kind of intermediate link between the very good RTX 2080 and the significantly more expensive RTX 2080 Ti. Our experts preferred the Super version over the Ti, since, all other things being equal and at a lower price, this modification is able to provide comfortable gaming at maximum speed in 4K 3840 × 2160 resolution, provided, of course, that the rest of the key hardware is powerful enough.

The TU104-450-A1 graphics processor is based on the 12nm process technology. It operates at a frequency from 1650 to 1965 – it is within these limits that factory overclocking is available. Video memory GDDR6 with an effective frequency of 15.5 GHz in a volume of 8 GB. Bus width – 356 bit. Supports high-speed data transmission using NVLink technology. Processor mathematics parameters: 3072 arithmetic logic units; shaders version 6.1; 192 texture units; 64 blocks of rasterization. On the firmware side, OpenGL version 4.6, DirectX 12, Vulkan, OpenCL 1.2 and CUDA version 7.5 are supported.

The card is noticeably larger than any of the cards discussed above – 328x140x57mm and takes up three slots. Requires additional power supply via two 8 pin + 8 pin connectors. Moreover, it requires a decent power supply, due to which the recommended power supply for this card in a gaming PC is 650W.

This model can provide exactly the same maximum resolution as any other in the review – 7680×4320. It also supports 4 simultaneously connected monitors. Some descriptions say support for five monitors, but this is not the case. The interfaces here are one HDMI2.0b, three DisplayPort1.4 and one USB Type-C. The USB-C port has a special purpose. This is the so-called VirtualLink for connecting virtual reality devices.

The cooling system consists of a large heatsink and three fans, efficient enough to handle 250W of heat dissipation. The maximum declared cooling noise is 24.9dBA.

Before buying, if the other components of the gaming PC are already available, it is imperative to compare the dimensions of the card with them. It will not be included in every building.

Advantages

• copes with the maximum graphics settings in games at 4K resolution;
• there is VirtualLink for virtual reality devices;
• support for NVLink technology;
• overclocking potential;
• effective cooling;
• Quiet enough work;
• attractive appearance.

disadvantages

• heavy, bends textolite of motherboards;
• squeak of chokes is heard under load.

GIGABYTE GeForce RTX 2080 SUPER 1860MHz 8192MB 15500MHz AORUS

Rating: 4.8

Rounding out the top three top-end gaming video cards at the beginning of 2020 according to the version is the actual analogue of the above model, but produced by another market leader – GIGABYTE.

From the platform's point of view, this card is a complete and absolute analogue of the previous one – with the same clock frequency, the same mathematics, and the same manufacturing process – 12nm. After all, exactly the same TU104-450-A1 graphics processor is installed here as in the above-described version. So we won't go into details here. This model was offered by our specialists only as an alternative to MSI for GIGABYTE fans or in case of accessibility problems – in this regard, these two cards will already act as alternatives to each other.

In general, in terms of electronics and firmware, these cards are complete twins. So, for example, the amount of memory is the same here – 8GB, and the parameters of the frequency and width of the bus are exactly the same. Exactly the same set of one-to-one ports, and the same standards are supported.

Some minor differences are observed in other points of the technical implementation. So, this model is slightly less overall – 287x115x50mm, but still occupies the same three slots. But perhaps these few centimeters on each side can be important for someone. The card needs additional power, only it is supplied not by 8 + 8pin connectors, but by 8 + 6pin. The power requirements for the power supply are the same – at least 650W. Finally, an important point – this model, being technically completely similar to the above-described MSI, is on average cheaper by as much as 10%.

Advantages

• all the advantages of the previous model;
• more compact than the previous model;
• on average at points of sale cheaper than the previous model.

disadvantages

• the same cons as the previous model.

---

Attention! This rating is subjective and does not constitute an advertisement and does not serve as a purchase guide. Before buying, you need to consult with a specialist.