Nvidia RTX DLSS: everything you need to know

By Jon Martindale, Jacob Roach and Kunal Khullar Updated April 16, 2025

Nvidia’s RTX series include some of the best graphics cards on the market and are known for two flagship features: real-time ray tracing and Deep Learning Super Sampling (DLSS). While ray tracing is no longer an exclusive feature as it is available on other GPUs as well as consoles like the PlayStation 5 and Xbox Series X, DLSS remains a bit more mysterious to many. At its core, DLSS uses AI-powered upscaling to render games at a lower resolution and intelligently reconstruct them to look like — or even better than — native 4K, significantly boosting performance without sacrificing visual fidelity.

With the arrival of DLSS 4 and the latest RTX 50-series GPUs, the technology has evolved even further, delivering smoother gameplay, better image quality, and smarter frame generation. It’s a win-win: higher frame rates, crisper visuals, and less strain on your GPU.

But there’s more going on under the hood than just upscaling. Here’s everything you need to know about DLSS, how it works, and why it’s become a must-have feature for modern PC gaming.

DLSS stands for deep learning supersampling. The “supersampling” bit refers to an anti-aliasing method that smooths the jagged edges that show up on rendered graphics. Over other forms of anti-aliasing, though, SSAA (supersampling anti-aliasing) works by rendering the image at a much higher resolution and using that data to fill in the gaps at the native resolution.

The “deep learning” part is Nvidia’s secret sauce. Using the power of machine learning, Nvidia can train AI models with high-resolution scans. Then, the anti-aliasing method can use the AI model to fill in the missing information. This is important, as SSAA usually requires you to render the higher-resolution image locally. Nvidia does it offline, away from your computer, providing the benefits of supersampling without the computing overhead.

This all works thanks to Nvidia’s Tensor Cores, found exclusively in RTX graphics cards (aside from enterprise-grade solutions like the Nvidia A100). Tensor Cores are purpose-built to accelerate the kinds of complex matrix operations and mixed-precision calculations used in AI and deep learning. By leveraging lower-precision formats like FP16 while preserving accuracy, they dramatically speed up tasks like neural network training and inference — and in gaming, that power is harnessed to drive features like DLSS.

Nvidia’s latest RTX 50-series graphics cards introduce fifth-generation Tensor Cores, pushing DLSS performance even further. With the addition of DLSS Multi-Frame Generation, these GPUs can now use AI to generate up to three extra frames for every traditionally rendered one — delivering a significant FPS boost without compromising visual quality.

There is no doubt that Nvidia is leading the charge in this area, though AMD’s FidelityFX Super Resolution feature is slowing trying to catch up. Even Intel has its own supersampling technology called Intel XeSS, or Intel Xe Super Sampling.

DLSS is the result of an exhaustive process of teaching Nvidia’s AI algorithm to generate better-looking games. After rendering the game at a lower resolution, DLSS infers information from its knowledge base of super-resolution image training to generate an image that still looks like it was running at a higher resolution. The idea is to make games rendered at 1440p look like they’re running at 4K or 1080p games to look like 1440p. DLSS 2.0 offered four times the resolution, allowing you to render games at 1080p while outputting them at 4K.

More traditional super-resolution techniques can lead to artifacts and bugs in the eventual picture, but DLSS is designed to work with those errors to generate an even better-looking image. In the right circumstances, it can deliver substantial performance uplifts without affecting the look and feel of a game; on the contrary, it can make the game look even better.

With the introduction of DLSS 3 on RTX 40-series of graphics cards, the frame rate gains were even more substantial. Nvidia not only used Tensor cores to make frames look better, but frames could be rendered using just AI. Additionally DLSS 3 delivered a performance boost by generating entire new frames between rendered frames, improving frame rates and responsiveness in games, especially on GPU- and CPU-limited systems.

DLSS 4 is a step up from DLSS 3, mainly thanks to Multi Frame Generation, which boosts performance by using AI to create multiple frames for every one the GPU renders. Second, it uses a new Transformer-based AI model that delivers clearer, more detailed images. DLSS 4 also improves existing features like Ray Reconstruction, making lighting effects look sharper and textures more accurate.

DLSS forces a game to render at a lower resolution (typically 1440p) and then uses its trained AI algorithm to infer what it would look like if it were rendered at a higher one (typically 4K). It does this by utilizing some anti-aliasing effects (likely Nvidia’s own TAA) and some automated sharpening. Visual artifacts that wouldn’t be present at higher resolutions are also ironed out and even used to infer the details that should be present in an image.

As Eurogamer explains, the AI algorithm is trained to look at certain games at extremely high resolutions (supposedly 64x supersampling) and is distilled down to something just a few megabytes in size before being added to the latest Nvidia driver releases and made accessible to gamers all over the world.

In effect, DLSS is a real-time version of Nvidia’s screenshot-enhancing Ansel technology. It renders the image at a lower resolution to provide a performance boost, then applies various effects to deliver a relatively comparable overall effect to raising the resolution.

The result can be a mixed bag, but in general, it leads to higher frame rates without a substantial loss in visual fidelity. Nvidia claims frame rates can improve by as much as 75% in Remedy Entertainment’s Control when using both DLSS and ray tracing. It’s usually less pronounced than that, and not everyone is a fan of the eventual look of a DLSS game, but the option is certainly there for those who want to beautify their games without the cost of running at a higher resolution.

The journey began with DLSS 1.0, which launched in 2019. It used a per-game trained AI model that rendered games at a lower resolution and then upscaled them to a higher one. However, this version had some major drawbacks. Because each game required specific training data, adoption was limited and inconsistent. Image quality also varied widely across titles, with many players reporting blur and ghosting artifacts that degraded the experience.

With the introduction of DLSS 2.0 Nvidia moved to a generalized AI model that no longer required game-specific training. This version introduced temporal feedback, meaning it used data from previous frames and motion vectors to reconstruct details more accurately. DLSS 2.0 included multiple modes—Quality, Balanced, Performance, and Ultra Performance—allowing users to choose between higher image fidelity or better frame rates.

DLSS 3, introduced in 2022 with the RTX 40 series, added another layer: Frame Generation. Unlike previous versions that only upscaled images, DLSS 3 could now generate entirely new frames using AI. This was made possible by an optical flow accelerator that examined motion between two frames and predicted what an intermediate frame should look like. This technology allowed for noticeable performance boosts, especially in CPU-bound scenarios, and was exclusive to the RTX 40 series and newer. While frame generation can introduce some latency, Nvidia included Reflex technology to minimize input lag and keep games responsive. Essentially, DLSS 3 is a combination of DLSS 2’s Super Resolution and the new AI-powered Frame Generation.

Back in January, Nvidia announced DLSS 4 — the latest iteration of the technology that will be available to the owners of RTX 50-series graphics cards. As mentioned, the biggest change this time around is the introduction of Multi Frame Generation which brings three additional AI-generated frames, giving a significant boost to frame rates on supported games.

There are currently six GPUs that support DLSS 4:

In our RTX 5090 review, we found that DLSS 4 brings a number of meaningful upgrades to Nvidia’s RTX 50-series GPUs. One of the most notable changes is the shift from a convolutional neural network (CNN) to a Transformer-based AI model. This update improves how the GPU reconstructs frames, leading to more accurate details and cleaner image quality — especially in fast-moving scenes where earlier versions of DLSS sometimes struggled.

Another key feature is DLSS Multi-Frame Generation (MFG), which expands on the frame generation introduced in DLSS 3. Instead of generating just one additional frame, DLSS 4 can now generate up to three AI frames for every one frame the GPU renders. This results in a significant performance boost, though image quality still depends heavily on the specific game implementation. DLSS 4’s multi-frame generation relies on a technique called frame interpolation, the same approach used in DLSS 3 as well as other frame generation tools like Lossless Scaling and AMD’s FSR 3.

In testing, Cyberpunk 2077 saw one of the biggest improvements, with performance nearly quadrupling when DLSS 4 was enabled. Other games like Alan Wake 2, Hogwarts Legacy, and Microsoft Flight Simulator also showed large gains in frame rate, particularly when ray tracing was turned on. However, it is important to note that the actual experience can vary — in some titles, the visual clarity of DLSS 4 was excellent, while others still showed artifacts or soft edges in motion.

Overall, DLSS 4 isn’t just a minor update — it brings real performance gains, especially in games that already support advanced ray tracing features. But as with past versions, the quality of results still depends on how well individual games implement the technology.

DLSS 4 is still new and is slowly making its way into more games. Here are the titles that currently support DLSS 4:

Refer to Nvidia’s DLSS 4 games list to see the latest compatible titles.

AMD is Nvidia’s biggest competitor when it comes to graphics technology. To compete with DLSS, AMD FidelityFX Super Resolution (FSR) was released in 2021. Although it achieves the same goal of improving visuals while raising frame rates, FSR works quite differently from DLSS. FSR renders frames at a lower resolution and then uses an open-source spatial upscaling algorithm to make the game look like it is running at a higher resolution and doesn’t factor in motion vector data. DLSS uses an AI algorithm to deliver the same results, but this technique is only supported by Nvidia’s own RTX GPUs. FSR, on the other hand, can work on just about any GPU.

In addition to FSR, AMD also has Radeon Super Resolution (RSR), which is a spatial upscaling technique that makes use of AI. While this sounds similar to DLSS, there are differences. RSR is built using the same algorithm as FidelityFX Super Resolution (FSR) and is a driver-based feature that is delivered via AMD’s Adrenalin software. RSR aims to fill the gap where FSR is not available, as the latter has to be implemented right into specific games. Essentially, RSR should work in almost any game, as it doesn’t require developers to implement it. Notably, FSR is available across newer Nvidia and AMD GPUs, and RSR, on the other hand, is only compatible with AMD’s RDNA cards, which include the Radeon RX 5000 and above.

FSR continues to evolve, with ​FidelityFX Super Resolution 4 (FSR 4) being the latest iteration of its upscaling technology. AMD is finally leveraging AI-powered frame generation which is claimed to significantly boost frame rates, with AMD claiming up to a 3.7x performance increase in supported titles. This version introduces advanced upscaling techniques and improved temporal stability, resulting in sharper visuals and smoother gameplay. FSR 4 is compatible with games that support FSR 3.1, although integration may require manual activation through driver settings. While initially developed for RDNA 4-based GPUs including the Radeon RX 9000 series, AMD has indicated plans to optimize FSR 4 for broader hardware compatibility in the future.

Intel has also been working on its own supersampling technology called Xe Super Sampling (XeSS), and unlike with FSR or DLSS, there are two different versions available. The first makes use of the XMX matrix math units, which are present in its new Arc Alchemist GPUs; these XMX units take care of all the AI processing on the hardware end. The other version makes use of the widely accepted four-element vector dot product (DP4a) instruction, thus removing the dependency from Intel’s own hardware and allowing XeSS to work on Nvidia and AMD GPUs.

Last year Intel introduced its Battlemage series of GPUs (B580 and B570) where it announced its updated XeSS 2 AI-powered upscaling technology designed to improve both gaming performance and visual quality. Building on the original XeSS, it introduced two major features: Frame Generation (XeSS-FG) and Xe Low Latency (XeLL). XeSS-FG boosts frame rates by generating additional AI-driven frames between rendered ones, while XeLL reduces input lag for a more responsive gaming experience. The core of XeSS 2, XeSS Super Resolution (XeSS-SR), uses AI to upscale lower-resolution frames to the display’s native resolution, similar to Nvidia’s DLSS and AMD’s FSR. These enhancements are especially effective on Intel Arc GPUs, positioning XeSS 2 as a direct competitor to DLSS 3 and FSR 3.