NVIDIA Reflex On vs On+Boost: Which Setting Is Best?
The Short Answer
For most players, On+Boost is the best setting — it delivers the lowest system latency of the three options. In typical LDAT-style measurements (reported from Fortnite in performance mode on an RTX 4070), system latency drops from roughly 22–30ms with Reflex off, to around 14–20ms with Reflex On, to about 10–15ms with On+Boost. The catch is small: On costs roughly 2% of your frame rate, and On+Boost costs around 4%, and Boost keeps your GPU clocks high, which means more power draw and heat. So the real answer depends on your GPU and whether your game is GPU-bound or CPU-bound — both covered below.
What NVIDIA Reflex Actually Does
Your PC doesn’t show you a frame the moment you move your mouse. The frame travels through a pipeline: input, game simulation, a render queue, the GPU, and finally the display. Latency builds up in that queue — if your GPU is busy rendering ahead, your inputs wait in line behind frames that were created earlier. Reflex attacks exactly that problem. It keeps the render queue just-in-time, so the GPU starts work on the freshest inputs instead of frames queued up in advance. Less queue means less delay between your hand and the screen. If input delay is a wider problem for you, our general guide to how to reduce input lag in pc games covers the rest of the chain, from monitors to USB polling.
The Latency Numbers
These are typical, reported figures from LDAT-style measurement (LDAT measures the full click-to-screen delay) in Fortnite’s performance mode on an RTX 4070. Your exact numbers will vary by game and system, but the pattern — and the size of the gaps — is consistent.
| Reflex setting | Typical system latency | Approx. FPS cost |
|---|---|---|
| Off | ~22–30ms | — |
| On | ~14–20ms | ~2% |
| On+Boost | ~10–15ms | ~4% |
Read the table the practical way: turning Reflex On is the big jump, cutting roughly a third off system latency. On+Boost then takes another meaningful bite — the difference between the bottom of the On range and the bottom of the On+Boost range is the difference between a good setup and a genuinely responsive one. Nobody should be playing a supported competitive game with Reflex Off.
What “Boost” Adds on Top of On
Boost changes how your GPU behaves while Reflex is active. Normally, a GPU lowers its clock speeds when it thinks the load is light — efficient, but clock changes introduce inconsistency. Boost holds the GPU’s clocks high and steady instead. Two results follow:
- Better 1% lows. Frame delivery stays more consistent, so the worst frames — the ones you actually feel as micro-stutter — improve.
- More power and heat. A GPU held at high clocks draws more power and runs hotter, even in menus and light scenes. On a well-cooled desktop this is a non-issue. On a laptop, it is a real consideration for fan noise and sustained thermals.
So Boost is not free latency. It trades a little efficiency for consistency, which is a trade most desktop players should happily make.
The Condition That Matters: GPU-Bound vs CPU-Bound
Reflex reduces latency created in the GPU render queue. That means it helps most when your GPU is the bottleneck — when your GPU usage sits at or near maximum while you play. In a GPU-bound game, the queue is long, and Reflex has a lot of delay to remove.
If your game is CPU-bound instead — an older processor, a simulation-heavy title, or settings so low that the GPU is half asleep — the queue is already short, and Reflex can only remove a little. You will still typically see some benefit from On, but Boost adds almost nothing in that situation, because the GPU was never the problem. Checking which side you are on takes ten seconds: watch GPU usage while playing. High and steady near maximum means GPU-bound; well below maximum with a maxed-out CPU means CPU-bound. This distinction is also why pairing Reflex with the wider best nvidia control panel settings for gaming matters less than most guides pretend.
Recommendation by GPU Tier
| Your setup | Recommended setting | Why |
|---|---|---|
| High-end desktop GPU (e.g. RTX 4070 class and up) | On+Boost | Lowest latency, ~4% FPS cost is easily absorbed, cooling handles the heat |
| Mid-range GPU, GPU-bound in your game | On+Boost | The latency gain is largest exactly here; switch to On only if thermals or noise bother you |
| Laptop GPU | On (try On+Boost and compare) | Boost’s constant high clocks add heat in a small chassis; test both in your longest session |
| Older GPU or CPU-bound games | On | Reflex still helps a little; Boost has almost nothing to fix |
One more practical rule: never leave Reflex Off in a supported title. Even in the worst case, On costs around 2% FPS — a trade almost every player should take for a visibly more responsive game.
How to Turn Reflex On
Reflex lives inside each game’s own settings — there is no single Windows or driver switch that enables it everywhere.
- Launch the game and open its settings menu.
- Go to the video or graphics section.
- Find the NVIDIA Reflex Low Latency option (usually near the resolution and rendering settings).
- Select On or On+Boost based on the recommendation table above.
- Restart the game if it asks — some titles only apply the change fully after a relaunch.
If the option is missing entirely, the game does not support Reflex, or your GPU is too old to offer it. Check for a game update first; support is sometimes added in patches long after launch.
FAQ
Does Reflex lower my FPS?
Yes, slightly. Typical reported figures are around 2% with Reflex On and around 4% with On+Boost. In exchange you get a much larger cut in system latency — usually the right trade in competitive games.
Should I use On+Boost on a laptop?
Try it, but judge it over a long session. Boost holds GPU clocks high, which raises heat and fan noise in a laptop’s small cooling system. If temperatures climb or performance drops later in a session, step back to plain On.
My game doesn’t have a Reflex option. Can I force it?
No — Reflex must be built into the game by its developer. Where it is missing, the closest alternatives are keeping GPU usage below maximum (so the queue stays short) and using your driver’s low-latency options, but neither fully replaces Reflex.
Conclusion
Turn Reflex On at minimum — that single change takes typical system latency from the 22–30ms range down to roughly 14–20ms. If you play on a desktop GPU with reasonable headroom, go one step further to On+Boost for the reported 10–15ms range and better 1% lows, accepting around 4% FPS and extra heat as the price. The only players who should hesitate are laptop users watching thermals and CPU-bound players, for whom plain On captures nearly everything Reflex has to give.
