ScreenSize.net finds most common desktop viewport is 1920x920, not 1920x1080

ScreenSize.net, a site that offers free browser viewport, device-pixel-ratio and screen-resolution checking tools, has published a report on the real-world dimensions of browser windows measured across its own site. The report draws on 763 total snapshot samples, with 12 excluded from the published rankings as dimension anomalies (viewports wider than 6,000 CSS pixels or taller than 4,000), and it states plainly that the numbers describe qualified ScreenSize.net sessions rather than global device or browser market-share estimates.

On desktop, drawn from 648 samples, the single most common viewport is 1920x920 CSS pixels, at 15.7% (102 samples). The full 1920x1080 pairing is only the second most common desktop pair, at 6% (39 samples). Looking at widths alone, 1920px leads at 30.7% (199 samples), followed by 1280px at 13.3% (86), 1520px at 9.7% (63), 800px and 2560px tied at 5.4% (35 each), 1480px at 5.2% (34), 1360px at 4.8% (31) and 1440px at 4.2% (27).

On mobile, drawn from 103 samples, the most common viewport width is 400px, at 64.1% (66 samples), well ahead of 420px at 20.4% (21), 380px at 9.7% (10), and 600px and 800px tied at 2.9% (3 each). By exact width and height together, 400x660 is the single most common pair at 31.1% (32 samples), a smaller share than the width figure alone because that same 400px width splits across several different heights, including 400x700 (9.7%), 400x720 (8.7%) and 400x780 (4.9%). Other widths also rank near the top: 420x800 (7.8%), 380x620 (6.8%), and 420x720 and 420x780 tied at 4.9% each.

Across desktop and mobile combined, 751 samples had a measured device pixel ratio (everything other than the 12 excluded anomalies), and a ratio of 1.00x is the most common bucket among them, at 64.6%.

The report documents its own methodology and limits. Legacy tablet samples are excluded because an earlier version of its classifier could mistake touch-enabled desktop computers for tablets, and viewports outside the 6,000 by 4,000 CSS pixel range are dropped from the published rankings, though they stay in aggregate storage for auditing. On privacy, a browser tab contributes at most one sample per session, results are rounded into monthly aggregate buckets, and the report states that no IP addresses, user IDs, precise timestamps or full user-agent strings are stored; Global Privacy Control and Do Not Track signals are honored. The snapshot was captured on 2026-09-14.

Key facts

The most common real-world desktop browser viewport is 1920x920 CSS pixels, 15.7% of 648 desktop samples (102 samples), per ScreenSize.net's own measurements.
The full 1920x1080 pairing is only the second most common desktop viewport, at 6% (39 of 648 samples), well behind 1920x920.
The most common mobile viewport width is 400px, 64.1% of 103 mobile samples (66 samples); the most common exact mobile pair is 400x660, at 31.1% (32 samples).
A device pixel ratio of 1.00x is the most common bucket across both desktop and mobile, at 64.6% of 751 measured samples.
The report is based on 763 total snapshot samples (12 excluded as dimension anomalies) from ScreenSize.net's own visitors, and it explicitly states the numbers are not global device or browser market-share estimates.

Why it matters

The report puts a specific, checkable number on something web developers usually only estimate: how tall a desktop browser window actually is in practice. Among 648 real desktop sessions, the single most common viewport is 1920x920 CSS pixels (15.7%), while the full 1920x1080 pairing is only the second most common pair, at 6% (39 of 648). Design decisions such as where content sits above the fold, or how tall a hero section can safely be, rest on assumptions about available viewport height; this dataset offers a concrete, if narrow, data point for that assumption instead of a guess.

Who it affects

Front-end and web developers who plan layouts, breakpoints and above-the-fold content are the natural audience for this kind of number, along with anyone running visual QA against assumed screen sizes. The sample itself also defines a narrower who: it is drawn only from visitors to ScreenSize.net's own viewport, device-pixel-ratio and resolution-checker tools, a population likely skewed toward developers and testers rather than a cross-section of everyone who browses the web.

How to use it

The report is a free, publicly readable page, part of the same ScreenSize.net site that hosts a live Viewport Size Checker, a Device Pixel Ratio Checker, a Screen Resolution Checker and a reference list of common screen resolutions, so a developer can check their own browser's numbers against the aggregate figures directly. As a planning reference, the data points toward a primary desktop breakpoint near 1920px wide with roughly 920px of vertical space rather than a full 1080px, mobile layouts built around a 400px width with heights commonly falling between about 660px and 780px, and the majority of sessions, 64.6% of 751 measured samples, on a 1.00x device pixel ratio rather than a high-DPR screen. The report itself cautions against treating any of this as global device or browser market-share data.

How solid is it

The sample is modest and fully self-selected. In total 763 snapshots were captured, of which 12 were excluded from the published rankings as dimension anomalies (wider than 6,000 CSS pixels or taller than 4,000), leaving 648 desktop and 103 mobile samples behind the headline percentages, and 751 samples with a measured device pixel ratio. The report documents real quality controls: a browser tab contributes at most one sample per session, figures are rounded into monthly aggregate buckets, and legacy tablet samples are dropped because an earlier version of the classifier could mistake touch-enabled desktop computers for them. What it does not give is a breakdown by browser, operating system or country, any confidence interval around the percentages, or a stated length or start date for the reporting period, only a single snapshot-captured date of 2026-09-14. The report is explicit that its numbers describe its own qualified sessions, not global device or browser market share.

Risks and caveats

The easiest misreading is to treat 1920x920 as a claim about physical monitor resolution. The report measures CSS viewport pixels, the space actually available to a webpage, not a display's native resolution, and it gives no per-sample device pixel ratio, so the text alone cannot confirm how many of those 1920-wide windows sit on a 1920x1080 monitor rather than a larger or scaled display. The sampling method, drawing only from visitors to a viewport and DPR-checking tool, plausibly skews toward developers and testers, and the report does not attempt to measure or correct for that skew. With no start date given for the reporting period, there is also no way to tell from the text whether these figures reflect a few days of traffic or a much longer stretch.

“These observations describe qualified ScreenSize.net sessions during the reporting period. They are not global device or browser market-share estimates.”

— ScreenSize.net viewport statistics report