Cacheable responses

The share of responses that carry cacheable headers.

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

At a glance the headline numbers for Cacheable responses

The share of responses that carry cacheable headers.

64.8%
on the typical site
half of sites sit at or below
84.2%
1 in 4 sites exceed this
the top quarter
98.1%
the heaviest 1%
the long tail
185,486
sites measured
desktop field data

64.8% of responses on the typical site are cacheable.

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

Distribution & median INP site count and median INP at each level of cacheable responses

0ms 63ms 125ms 188ms 250ms
200ms
0 92419 184837
0–0 1+
Good (≤200ms) Needs improvement Poor (>500ms) Site count
The State of Web Vitals · Q2 2026 · 189,915 sites · desktop field datacorewebvitals.io/state-of-cwv
3

Passing INP by cacheable responses which level passes the INP most often

Cacheable responsesSitesPassing INPINP
0–0 184,837 99% 49ms
1+ 649 99% 39ms
Good Needs Improvement Poor Faded rows: under 100 sites

Cacheable responses 64.8%. p75 84.2%. p99 98.1%. At the low end (0–0): INP 49ms. At the high end (1+): INP 39ms. 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

This is the share of responses a browser is allowed to keep. Everything outside that share gets re-downloaded on the next visit. Repeat visitors are in your field data too, and for them cacheability is the difference between reading from disk and crossing the network.

Uncacheable static assets are almost always an accident: a missing header on a font, a no-cache default on an image bucket. If a file has a hash in its name, there is no reason it should ever be fetched twice.

Related signals HTML compression → Cache strategy mix → HTTP protocol mix → HTTP protocol → Chrome field data from 189,915 sites, representing millions of real page loads · How we measured