DLSS does not have one single effect on input lag. DLSS Super Resolution usually reduces system latency when it raises your real rendered frame rate in a GPU-limited game. DLSS Frame Generation is different: it inserts AI-generated frames between rendered frames, so the displayed FPS rises without your game processing input at that same new rate. NVIDIA Reflex is used alongside Frame Generation to keep latency under control.
The short version: Super Resolution can make a game more responsive; Frame Generation makes motion look smoother but should not be treated as equivalent to the same FPS rendered normally.
DLSS latency at a glance
| Feature | What it does | Typical latency effect | Best use |
|---|---|---|---|
| DLSS Super Resolution | Renders at a lower internal resolution and reconstructs the output | Often lowers latency if it increases rendered FPS | GPU-limited games, high resolutions, ray tracing |
| DLSS Frame Generation | Creates extra frames between rendered frames | Adds processing versus the same base render path; Reflex helps offset it | Smoother single-player/high-fidelity gaming |
| DLSS Multi Frame Generation | Generates multiple displayed frames per rendered frame on supported hardware | Very high displayed FPS does not mean equally high input-sampling rate | High-refresh visual smoothness when base FPS is already healthy |
| NVIDIA Reflex | Synchronizes CPU/GPU work to reduce the render queue and system latency | Reduces latency | Enable when the game supports it, especially with Frame Generation |
Why Super Resolution can reduce input lag
DLSS Super Resolution reduces the GPU’s rendering workload by producing a lower-resolution frame and using AI to reconstruct the higher-resolution output. NVIDIA describes Super Resolution as a way to boost performance while maintaining image quality.
If your graphics card is the bottleneck, less work per frame can mean a higher rendered frame rate. That matters because shorter frame times generally improve responsiveness. At 60 FPS, one frame takes about 16.7ms; at 120 FPS, it takes about 8.3ms.
That does not mean DLSS itself is magically “negative latency.” The reconstruction step has a cost. The reason total latency can fall is that the GPU may finish frames much faster than it did at native resolution.
This is why the answer depends on the game and bottleneck. If DLSS Quality takes you from 55 FPS to 90 FPS because you were GPU-bound, the system will usually feel more responsive. If you are already CPU-limited and DLSS barely changes the rendered FPS, the latency benefit may be tiny.
For a deeper look at the preset trade-offs, see our DLSS Quality vs Performance guide.
Frame Generation is a different latency problem
DLSS Frame Generation creates new frames between frames rendered by the game. That can dramatically increase the number shown by an FPS counter, but generated frames do not represent a fresh game simulation and fresh input sample in the same way a newly rendered frame does.
So if a game renders 70 real frames per second and Frame Generation raises the displayed output far above that, the animation can look much smoother while the controls still feel closer to the underlying rendered frame rate than the displayed number suggests.
NVIDIA’s own developer guidance treats latency as something that must be validated separately when Frame Generation is integrated. The company specifically requires a correct NVIDIA Reflex integration and warns that an incorrect Reflex setup can produce a significant latency increase.
Does Frame Generation always make latency worse than native?
Not necessarily. The comparison matters. If “native” means an extremely demanding ray-traced mode running at a low frame rate and the DLSS configuration also uses Super Resolution plus Reflex, the complete DLSS pipeline can still end up with lower system latency than native max settings. NVIDIA has demonstrated cases where DLSS 3 with Reflex both increased performance and reduced latency compared with native rendering.
But if you compare the same underlying rendered FPS with and without Frame Generation, Frame Generation should not be interpreted as creating the responsiveness of a truly rendered higher frame rate. Its main benefit is visual smoothness.
What NVIDIA Reflex changes
NVIDIA Reflex is designed specifically to reduce system latency by coordinating CPU and GPU work more efficiently. A major part of that is reducing unnecessary render queueing, so the frame being displayed is based on more recent input.
Reflex is separate from Super Resolution. A game can support Reflex even when you are rendering at native resolution. With Frame Generation, however, Reflex becomes particularly important because it helps counter the extra latency introduced by the generated-frame pipeline.
NVIDIA’s DLSS 3 documentation says Frame Generation is integrated with Reflex for responsiveness, and its developer documentation recommends measuring latency with tools such as FrameView or the Reflex Latency Analyzer rather than assuming that a higher FPS number automatically means lower latency.
What about DLSS 3.5, DLSS 4 and newer versions?
Version numbers can make this topic confusing because different DLSS features evolve separately.
DLSS 3.5’s headline addition was Ray Reconstruction, not a new universal Frame Generation latency fix. Newer DLSS generations have continued to improve Super Resolution, Frame Generation and Multi Frame Generation, but the fundamental distinction remains: reconstruction/upscaling can increase real rendered FPS, while generated frames primarily increase displayed smoothness.
NVIDIA’s current DLSS technology overview separates Super Resolution, Frame Generation and Multi Frame Generation as distinct parts of the DLSS suite. Treat them separately when you evaluate latency.
When should you use DLSS for the lowest input lag?
If you play competitive shooters
Prioritize the highest stable rendered frame rate and enable Reflex when supported. DLSS Super Resolution can be useful when you are GPU-bound. Frame Generation is usually less attractive when absolute response time matters more than visual smoothness.
Also remember that refresh rate matters. A high rendered frame rate has more value when your display can show it efficiently, which is why our 75Hz vs 144Hz vs 240Hz guide looks at frame time as well as monitor Hz.
If you play demanding single-player games
Frame Generation can be an excellent trade. In a visually heavy RPG or ray-traced game, smoother camera motion can be more noticeable than a small difference in control latency. Start with a healthy base rendered frame rate, enable Reflex, and judge the result by feel rather than the FPS counter alone.
If you are CPU-bound
Super Resolution may do very little for latency because lowering GPU render resolution cannot make the CPU prepare game frames faster. Frame Generation can still raise displayed FPS in some CPU-limited situations, but again, that does not turn the underlying simulation into a genuinely higher-rate game loop.
DLSS mode does not determine latency by itself
It is tempting to say Performance mode always has the lowest lag because it usually produces the highest FPS. That is too simplistic.
Performance mode can lower latency more than Quality mode when the GPU is the limiting factor and the extra FPS is real. But if the game becomes CPU-limited, dropping the internal resolution further may stop helping. The best mode is the one that gives you the frame rate and image quality balance you actually need.
Our FSR vs native resolution guide covers the same broader idea from the AMD/upscaling side: headline FPS alone does not tell the whole story.
How to test DLSS latency on your own PC
- Pick a repeatable scene or built-in benchmark.
- Record native-resolution FPS and, if available, system latency.
- Enable DLSS Super Resolution without Frame Generation and retest.
- Enable Reflex and retest.
- If your GPU supports it, enable Frame Generation last and compare responsiveness as well as displayed FPS.
If the game or hardware supports NVIDIA latency measurement, use it. NVIDIA recommends FrameView or a Reflex Latency Analyzer for validating system-latency changes.
FAQ
Does normal DLSS increase input lag?
DLSS Super Resolution usually does not increase total system latency when it meaningfully raises your rendered frame rate. In a GPU-bound game, it can reduce latency. In a CPU-bound game where FPS barely changes, the benefit may be small.
Does DLSS Frame Generation increase input lag?
Frame Generation adds a generated-frame stage and should not be equated with the responsiveness of the same FPS rendered normally. NVIDIA pairs it with Reflex to minimize system latency, and a complete DLSS + Reflex setup can still beat a much slower native configuration.
Should I use Frame Generation for competitive gaming?
If your priority is the lowest possible latency, usually not. Favor a high native or Super Resolution rendered frame rate plus Reflex. Frame Generation is more compelling when visual smoothness is the priority.
Does a higher DLSS FPS number always mean lower latency?
No. With Super Resolution, higher FPS generally represents more rendered frames. With Frame Generation or Multi Frame Generation, some of the displayed frames are generated, so the FPS counter can rise much more than the game’s input-to-render update rate.
Bottom line
DLSS Super Resolution usually helps latency when it removes a GPU bottleneck. Frame Generation can improve smoothness without giving you the input response of the same fully rendered FPS. Use NVIDIA Reflex whenever the game supports it, especially alongside Frame Generation, and judge latency separately from the frame-rate counter.
