What Is Mouse Polling Rate and Why It Matters

Polling rate is how often a mouse tells the computer where it is. The mouse sends a position report at a fixed interval. A 1000 Hz mouse sends one report every millisecond. A 125 Hz mouse sends one every 8 milliseconds. The number is the frequency in hertz.

The standard classes you see on packaging are 125 Hz, 250 Hz, 500 Hz, 1000 Hz, 2000 Hz, 4000 Hz, and 8000 Hz. Each step halves the reporting interval. At 125 Hz the interval is 8 ms. At 1000 Hz it is 1 ms. At 8000 Hz it is 0.125 ms. The interval is the inverse of the frequency. These classes come from the USB HID specification for mice, which defines the report interval field that the device uses to tell the host how often it will send reports.

Higher polling rates reduce the time between a physical movement and the cursor updating on screen. That can feel snappier in fast games, especially in first-person shooters where micro-adjustments matter. The trade-off is CPU and USB bandwidth. A 1000 Hz mouse uses a small but measurable share of USB interrupt capacity. An 8000 Hz mouse uses roughly eight times as much. On modern systems the difference is rarely a bottleneck, but on older hardware or crowded USB controllers it can show up as stutter. Some mice also draw more battery power at higher rates when running wireless.

Polling rate is not DPI. DPI (dots per inch, often called CPI — counts per inch) is how many sensor counts the mouse registers per inch of physical travel. Polling rate is how often those counts are delivered. A 26,000 DPI mouse at 125 Hz still only reports 125 times per second. A 400 DPI mouse at 1000 Hz reports 1000 times per second. They are independent settings. Changing one does not change the other. You can think of DPI as the resolution of the sensor and polling rate as the refresh rate of the connection.

The display refresh rate interacts with polling rate in a way that surprises many buyers. A 144 Hz display refreshes every 6.9 ms. A 1000 Hz mouse reports every 1 ms. The browser, by default, aligns pointer event delivery to the display frame. This means that even with a 1000 Hz mouse, the browser may only receive updates at the display refresh rate unless the page reads coalesced events. The getCoalescedEvents method on PointerEvent returns the intermediate samples that the browser combined into a single frame-aligned event. This site uses that API to recover the higher-rate samples.

A browser-based test cannot read the mouse firmware or the USB packet stream. It receives pointer events after the operating system processes the reports. The browser aligns delivery to display frames by default. The number the test shows is what the browser observed — a lower bound, not a hardware measurement. The test on this site labels every result “browser-observed” for that reason.

If you change the polling rate in vendor software, the browser-observed rate will usually move in the same direction. But it will not match exactly. A 1000 Hz setting typically reads a little under 1000 Hz in the browser. A 4000 Hz or 8000 Hz setting reads far lower because current browsers cap the number of distinct timestamps they can deliver per second. See the browser guide for the mechanism.

The test also shows a stability graph. That graph plots the 1-second window rates across the session. It lets you see whether the rate holds steady or swings between drops and spikes. Steadiness matters more than the peak number for most use cases. A mouse that holds 950 Hz steadily is often more useful than one that spikes to 1000 Hz but drops to 600 Hz every few seconds. The variation percentage — coefficient of variation across windows — summarizes this in one number. Lower is steadier.

Keyboard polling rate cannot be measured from a browser. Key events arrive after OS processing, and the WebHID API does not expose keyboard input to web pages. For a detailed breakdown of why keyboard report intervals cannot be measured online, see our technical guide on what browser keyboard tests can and cannot show.

For a deeper look at why browser readings differ from hardware specs, see the browser guide page. For the full measurement method, see the how we measure page.