The State of Web Vitals

LCP image format

The image format of the LCP resource.

Field data PhoneDesktopAll Scope All sites Q2 2026 edition · Desktop field outcomes
Metric LCP INP CLS
1

At a glance the headline numbers for LCP image format

The image format of the LCP resource.

15
Categories
111,844 sites
49.6%
Top share
jpg
0.03
Top CLS
Median for jpg

49.6% of LCP images are JPEG. 26.6% are WebP, 6.1% AVIF.

The State of Web Vitals · Q2 2026 · 189,915 sites · desktop field datacorewebvitals.io/state-of-cwv
2

LCP image format who uses what, and how stable each group is

Median CLS
0
0.05
0.10
0.15
0.20
0.25
avif, image/avif0.000% of sites
*0.010% of sites
avif0.026% of sites
gif0.021% of sites
jpg0.0350% of sites
webp0.0327% of sites
png0.0315% of sites
svg+xml0.032% of sites
PNG0.030% of sites
pjpeg0.070% of sites
pipeg0.410% of sites
VariantShare of sitesMedian
avif, image/avif 1 sites
0%
0.00
* 104 sites
0%
0.01
avif 6,824 sites
6%
0.02
gif 651 sites
1%
0.02
jpg 55,475 sites
50%
0.03
webp 29,753 sites
27%
0.03
png 16,520 sites
15%
0.03
svg+xml 2,499 sites
2%
0.03
PNG 2 sites
0%
0.03
pjpeg 8 sites
0%
0.07
pipeg 1 sites
0%
0.41

Avif, image/avif sites post the best median CLS (0.00). Pipeg sites trail at 0.41. Correlation, not causation. computed

The State of Web Vitals · Q2 2026 · 189,915 sites · desktop field datacorewebvitals.io/state-of-cwv
3

Passing CLS per LCP image format which group passes the CLS most often

VariantSitesPassing CLSCLS
PNG 0% 2 100% 0.03
Avif, image/avif 0% 1 100% 0.00
Avif 6.1% 6,824 79% 0.02
Webp 26.6% 29,753 78% 0.03
Gif 0.6% 651 78% 0.02
Svg+xml 2.2% 2,499 77% 0.03
Jpg 49.6% 55,475 76% 0.03
Png 14.8% 16,520 76% 0.03
* 0.1% 104 68% 0.01
Pjpeg 0% 8 67% 0.07
Pipeg 0% 1 0.41
Bmp 0% 2 —
JPEG 0% 1 —
X-pict 0% 1 —
X-ms-bmp 0% 2 —
Good Needs Improvement Poor Sorted best-passing first · median colored by its own rating · pass = good CLS (0.1 at p75) · one value per site

79% of Avif sites pass CLS. * trails 11 points behind, leaving 32% of its sites failing. computed

The State of Web Vitals · Q2 2026 · 189,915 sites · desktop field datacorewebvitals.io/state-of-cwv
4

Why this matters for the Core Web Vitals, and where to start fixing it

The format sets the bytes per pixel, and the LCP's load time scales with the bytes. The same hero photo is far smaller as AVIF or WebP than as JPEG, and PNG for photos is the classic mistake: a lossless format doing a lossy job. SVG is its own case, tiny and sharp at any size for logos and illustrations.

Nobody should convert images by hand. An image CDN or a build step serves the best format each browser supports from one source file. The format mix is mostly a measure of which sites have that step.

Go fix it Optimize images for Core Web Vitals → JPEG XL and Core Web Vitals → Optimize the LCP image →
Related signals LCP priority hint → LCP origin → LCP responsive markup → LCP discovery → Chrome field data from 189,915 sites, representing millions of real page loads · How we measured