The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →In a September 2026 re-probe of 1,000 YouTube videos, WebP maxresdefault thumbnails at 1280×720 used about half the bytes of their JPEG counterparts. At 480×360, the result was less consistent: WebP was larger in about one-third of all paired files, and in 72% of the sampled 2026 uploads. Those are file-size comparisons, not image-quality results.
Are YouTube WebP thumbnails smaller than JPEG?
Usually at the 1280×720 maxresdefault tier in the measured sample; not reliably at 480×360. Thumbnails Grabber compared public YouTube thumbnail URLs for 1,000 videos, reporting medians only for videos where both formats were available.
As an Amazon Associate I earn from qualifying purchases.
| Thumbnail variant | Dimensions stated in the study | Median JPEG | Median WebP | Median per-video WebP/JPEG ratio | Pairs where WebP was larger |
|---|---|---|---|---|---|
maxresdefault |
1280×720 | 96.7 KB | 50.8 KB | 0.53 | 10 of 965 (1.0%) |
sddefault |
640×480 | 41.3 KB | 23.3 KB | 0.55 | 14 of 970 (1.4%) |
hqdefault |
480×360 | 19.0 KB | 15.6 KB | 0.75 | 324 of 970 (33.4%) |
The ratio is the median of individual videos’ WebP-to-JPEG ratios; it is not calculated by dividing the two median sizes. Across the 965 paired 1280×720 files, total WebP bytes were 59.1% of total JPEG bytes. At 640×480, WebP bytes totaled 59.5% of JPEG bytes across the paired files. Thumbnails Grabber’s 2026 study reports these as sample measurements, not a census of YouTube.
Why can a 480×360 WebP thumbnail be bigger than JPG?
The 480×360 result depends strongly on which videos are compared. In the study’s older-skewing group of Wikidata-referenced videos, the median ratio was 0.65 and WebP was larger in 10.4% of pairs. For recent uploads, the median ratio was 1.06 and WebP was larger in 56.1%. Among 217 sampled uploads from 2026, the median ratio was 1.13 and WebP was larger in 72%.
Vertical Shorts in the 480×360 subset showed another sharp difference: WebP was larger in 85.3% of 116 pairs. Yet for the same Shorts subset’s max-resolution files, the median WebP/JPEG ratio was 0.40. These results are a reason not to treat one thumbnail tier as a proxy for another. The study observed the increase in ratios for small thumbnails from recent uploads but did not establish its cause.
What the measurements do—and do not—say
They compare transferred file bytes, not visual fidelity. YouTube’s encoder settings are unpublished, and the formats are separate encodes. The study did not assess whether one looked sharper, had fewer artifacts, or was otherwise better. A smaller file therefore does not establish equal or superior image quality.
The sample began with 9,055 public videos across two samples. The format re-probe selected 500 recent uploads and 500 Wikidata-referenced videos, all with a JPEG maxresdefault. JPEG sizes were measured on 7 September 2026 and WebP sizes on 19 September 2026; the study reports 3,000 requests and zero errors. Because the measurements were twelve days apart, a thumbnail replacement between requests could mean the two files represent different frames. The study also did not decode WebP dimensions, so the stated tier dimensions should not be read as an independent verification of every WebP file.
These figures describe publicly accessible files at those measurement dates. CDN availability and byte sizes may change when thumbnails are updated or reprocessed. A separate 300-video study fetched on 31 August 2026 also found smaller median WebP files at max resolution—60,345 bytes versus 141,111 bytes for JPEG—but its curated IDs were not randomly selected from all uploads. It supports the direction of the max-resolution finding, not a universal compression guarantee. The 300-video study gives its sample and methods.
Rank #3
What this means for developers and creators
For a page loading a YouTube thumbnail
Use the actual variant and video cohort relevant to your page when estimating bandwidth. The 1280×720 sample strongly favored WebP by bytes, while 480×360 results varied by upload era and format pairing. If you can choose the delivered format, compare the available files rather than assuming WebP is always smaller; include a fallback for clients or delivery paths that do not support WebP.
Do not choose a format on byte count alone when visual quality matters. Compare the actual images at the rendered size and assess artifacts, crop, and clarity separately. The study provides no quality scoring to substitute for that check.
For creators uploading custom thumbnails
The measured files are public CDN derivatives, not the original image creators upload. YouTube Help recommends JPG or PNG for custom thumbnail uploads and says, “Your custom thumbnail image should be as large as possible.” Its upload guidance is separate from the formats and byte sizes returned by public thumbnail URLs. YouTube Help: Add custom thumbnails on YouTube.
Why format names alone do not predict thumbnail size
WebP supports lossy and lossless compression, transparency, metadata, color profiles, and animation. RFC 9649 describes it as optimized for fast image transfer over networks, but those capabilities do not determine how a particular platform’s encodes compare. YouTube’s separate encoding choices and the thumbnail content both matter. RFC 9649 defines the format.
Google’s 2010 WebP comparison reported 41.30% average compression under its non-negative-compression-gain method, but it covered the first released WebP version and explicitly says a newer study supersedes it. It is historical context, not evidence for YouTube’s current thumbnail sizes. Google’s comparative study.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
