Cyberpunk 2077 finally gets DLSS 3 patch & boosts performance by up to 45%
CD Projekt Red’s Cyberpunk 2077 will officially get its DLSS 3 patch, offering much-improved performance when played on a 40-series Nvidia GPU.
Despite already being tested in our 4090, 4080, and 4070 Ti reviews, Cyberpunk 2077‘s DLSS 3 patch is set to be launched today.
Cyberpunk 2077 was used in a majority of the promotional materials by Nvidia, showing off not only the FPS bump that it could achieve but also the new frame generation technology being implemented.
Frame generation allows the 40-series card to effectively ‘look ahead’ at the motion on the screen and then generate the next frame instead of the game itself, easing up the GPU’s resources.
Cyberpunk DLSS 3 performance (4K)
| Game | RTX 4070 Ti | RTX 4080 | RTX 4090 |
| Cyberpunk 2077 (Ultra, RT Ultra, DLSS Performance, Frame Generation) | 86 FPS | 108 FPS | 116 FPS |
| Cyberpunk 2077 (Ultra, RT Ultra, DLSS Performance) | 59 FPS | 74 FPS | 90 FPS |
| Cyberpunk 2077 (Ultra, RT Ultra, DLSS off) | 29 FPS | 29 FPS | 37 FPS |
In our testing with the 4090, we found that using frame generation allowed us to hit 116FPS at Ultra and use the performance setting on DLSS. With frame generation off, we saw a dip to 90FPS. It is a staggering increase over the performance achieved at a native resolution. Between a few of the cards tested when compared to DLSS 2, you could get up to 45% more performance.
This patch doesn’t introduce the touted Overdrive Mode, which is set to improve how ray tracing works within the world on Nvidia 40-series GPUs.
The global art director for CD Projekt Red, Jakub Knapik said:
“Playing Cyberpunk 2077 with the AI-powered performance of NVIDIA DLSS 3 is the best way to access the stunning visuals the game has to offer — this technology helps us take a new step toward combining unconstrained visual fidelity with exceptional performance.“
What is DLSS?
DLSS, or Deep Learning Super Sampling, is an evolving technology that Nvidia implement to get more out of their cards. It works by lowering the resolution of the game and then using AI and machine learning to blow it back up to the user’s desired resolution. This, in turn, takes a load off of the GPU and allows the player to experience the game without hassle.