Core Web Vitals Thresholds (2026): Good LCP, INP and CLS Scores Explained

Official Core Web Vitals thresholds for LCP, INP and CLS (good, needs improvement, poor), how the 75th percentile works, field vs lab data, INP replacing FID, and what Google says about rankings.

Vehicle dashboard with a speedometer and tachometer side by side
Vehicle dashboard with a speedometer and tachometer side by side — Photo: Spacekid / Wikimedia Commons, CC0

Open Search Console or PageSpeed Insights and you get a wall of green, orange and red labels for LCP, INP and CLS. The labels are easy to read, but the rules behind them are not obvious: which exact number flips a page from "good" to "needs improvement," why your lab score can be 95 while the page still fails, and whether any of this actually moves rankings.

This guide answers those questions using only Google's own documentation: web.dev, Google Search Central, the PageSpeed Insights docs and Search Console Help, all checked on November 13, 2026. t-trends did not measure any particular site for this article. Every number below is an official threshold or an official recommendation, linked to the page it came from.

Core Web Vitals thresholds at a glance

Metric What it measures Good Needs improvement Poor
LCP (Largest Contentful Paint) Loading: render time of the largest image, text block or video in the viewport 2.5 s or less Over 2.5 s up to 4 s Over 4 s
INP (Interaction to Next Paint) Responsiveness: delay from a click, tap or key press to the next frame 200 ms or less Over 200 ms up to 500 ms Over 500 ms
CLS (Cumulative Layout Shift) Visual stability: the largest burst of unexpected layout shifts 0.1 or less Over 0.1 up to 0.25 Over 0.25

Basis: the threshold tables in Search Console Help and About PageSpeed Insights, checked 2026-11-13. Each metric is judged at the 75th percentile of page loads, with mobile and desktop rated separately. web.dev lists all three metrics as "stable" in its Web Vitals lifecycle (experimental, pending, stable).

Upper limit of the "Good" range in PageSpeed Insights (milliseconds)LCP and INP are Core Web Vitals; FCP and TTFB are supporting metrics (PageSpeed Insights labels TTFB experimental). CLS has no unit and is not shown. Source: About PageSpeed Insights, checked 2026-11-13
  • LCP (Largest Contentful Paint)2,500 ms
  • FCP (First Contentful Paint)1,800 ms
  • TTFB (Time to First Byte)800 ms
  • INP (Interaction to Next Paint)200 ms

PageSpeed Insights also reports two supporting loading metrics. First Contentful Paint is good at 1,800 ms or less and poor above 3,000 ms. Time to First Byte, which PageSpeed Insights labels experimental, is good at 800 ms or less and poor above 1,800 ms. web.dev describes both as helpful for diagnosing LCP problems, but neither is a Core Web Vital.

How the 75th percentile works

web.dev recommends measuring "the 75th percentile of page loads, segmented across mobile and desktop devices." Picture 100 visits sorted from fastest to slowest: the 75th visit has to land inside the good range. Because it is not an average, a few very fast visits cannot hide a slow tail. If more than a quarter of visits on slower phones or networks miss the threshold, the metric is not rated good.

The PageSpeed Insights docs spell out when a page or origin passes the Core Web Vitals assessment:

Data available Assessment result
Enough data for LCP, INP and CLS Passes if all three 75th percentiles are Good
INP lacks sufficient data Passes if the 75th percentiles of LCP and CLS are both Good

Search Console uses grouping on top of that. Its report clusters URLs with similar experiences, and a group's status "defaults to the slowest status assigned to it for that device type." So if LCP and INP are good but CLS needs improvement, the whole group shows as needs improvement. The same group can be good on desktop and poor on mobile.

Field data vs lab data

Most confusion about Core Web Vitals comes from mixing two kinds of data.

Field data Lab data
Source Chrome User Experience Report (CrUX): anonymized metrics from real Chrome users Lighthouse, running the page in a simulated environment
Time window The previous 28-day collection period The moment you run the test
Where you see it Top of PageSpeed Insights; Search Console Core Web Vitals report PageSpeed Insights performance score; Lighthouse
Used for Core Web Vitals pass or fail Debugging causes
Limitation Low-traffic pages may have no data Cannot reflect your real visitors' devices and networks

The Lighthouse performance score uses its own scale: per the PageSpeed Insights docs, 90 or above is good, 50 to 89 needs improvement, and below 50 is poor. That score is not the Core Web Vitals assessment. A page can score in the 90s in the lab and still fail in the field if many real visitors use slow phones.

New and small sites often have no CrUX data at all. Search Console Help says that when a URL group lacks data, the report creates "a higher-level origin group" instead, and a URL group without threshold data for both LCP and CLS does not appear in the report. Until field data shows up, the lab section of PageSpeed Insights is the practical way to find problems.

INP replaced FID on March 12, 2024

Date What changed Source
May 2023 Google announced INP would become a Core Web Vital in March 2024 web.dev blog
March 12, 2024 INP officially became a Core Web Vital and replaced First Input Delay; FID was removed from Search Console the same day web.dev: INP becomes a Core Web Vital on March 12
September 9, 2024 Deadline for applications using the CrUX and PageSpeed Insights APIs to move from FID to INP web.dev: INP is officially a Core Web Vital

The switch matters because the two metrics are not equivalent. According to web.dev's INP article, INP looks at mouse clicks, touchscreen taps and key presses across the whole visit (hovering, scrolling and zooming are not counted). For most pages the worst interaction is reported; on pages with many interactions, one highest value is ignored for every 50 interactions to filter out outliers. Each interaction is split into input delay, processing duration and presentation delay. If an older guide quotes an "FID under 100 ms" target, that metric is no longer part of Core Web Vitals.

Do Core Web Vitals affect ranking?

Here is what Google's Search Central pages actually say, quoted from Understanding page experience:

  • "Core Web Vitals are used by our ranking systems."
  • "There is no single signal."
  • "Other page experience aspects don't directly help your website rank higher in search results."
  • "Google Search always seeks to show the most relevant content, even if the page experience is sub-par."
  • "Trying to get a perfect score just for SEO reasons may not be the best use of your time."

The Core Web Vitals page adds: "We highly recommend site owners achieve good Core Web Vitals for success with Search." Read together, the message is consistent. Getting into the good range is worth doing and aligns with what ranking systems reward, but it does not replace relevant, useful content. Google publishes no figure for how much a better score moves a page, so any article promising a specific ranking gain is not quoting Google.

Fix checklist by metric (from web.dev)

The items below come from web.dev's most effective ways to improve Core Web Vitals and its per-metric optimization guides. Which ones matter for your site depends on what PageSpeed Insights diagnostics flag, so start there.

LCP checklist

web.dev's Optimize LCP guide splits LCP into four subparts and suggests roughly 40% each for time to first byte and resource load duration, with resource load delay and element render delay each under 10%. It calls these guidelines, not strict rules.

web.dev's suggested share of LCP time by subpart (%)Approximate guidance from web.dev's Optimize LCP article. Resource load delay and element render delay are recommended to stay under 10%, so 10 is shown as the ceiling. web.dev calls these guidelines, not strict rules. Checked 2026-11-13
  • Time to first byte40%
  • Resource load delay10%
  • Resource load duration40%
  • Element render delay10%
  • Make the LCP image discoverable in the initial HTML (src, srcset or <link rel="preload">), not hidden behind a data-src attribute that needs JavaScript
  • Add fetchpriority="high" to the LCP image
  • Never lazy-load the LCP image; keep loading="lazy" for images below the fold
  • Serve HTML from a CDN to cut TTFB
  • Consider instant navigations with bfcache and prerendering via the Speculation Rules API

Static sites start with an advantage on TTFB because there is no server-side rendering per request. If you are choosing a host for one, see the official free-tier limits in Free Static Site Hosting Compared. For shared hosting, check what you will pay after the intro term in Web Hosting Renewal Prices Compared.

INP checklist

  • Break up long tasks: web.dev defines a long task as anything over 50 ms on the main thread
  • Yield often, for example with scheduler.yield() where the browser supports it
  • Remove unnecessary JavaScript: find unused code with the Coverage tool, split code, and remove unused tags from your tag manager
  • Avoid large rendering updates: avoid forced layout and layout thrashing, keep the DOM small, and consider content-visibility for off-screen content

CLS checklist

  • Set width and height on images, use aspect-ratio for other embeds, and reserve a sensible min-height for ads and dynamic content
  • Keep pages eligible for the back/forward cache; web.dev notes pages with Cache-Control: no-store or unload listeners are often ineligible
  • Animate with transform (such as translateX) instead of top, left, margin or border

Remember how CLS is scored: per web.dev's CLS article, each shift scores impact fraction times distance fraction, shifts are grouped into session windows "with less than 1-second in between each shift and a maximum of 5 seconds for the total window duration," and shifts within 500 ms of user input are excluded. The usual culprit is an ad or image that loads late and pushes text down while someone is reading.

A practical order of work and caveats

  1. Start in Search Console. It shows field data grouped by URL pattern, so you see what real visitors experience. Check mobile and desktop separately.
  2. Test a representative URL in PageSpeed Insights. The field section tells you which metric fails; the lab diagnostics point to causes.
  3. Validate after fixing. In the Core Web Vitals report, "Start Tracking" begins a 28-day monitoring session.

Keep these limits in mind:

  • Field data covers a rolling 28-day period, so fixes show up gradually, not the next day.
  • INP only exists when people interact. Article pages with few clicks may lack INP data and be assessed on LCP and CLS alone.
  • Adding ad code can hurt CLS and INP at the same time. Reserving ad space up front is what web.dev recommends.

Related reading

FAQ

What is a good INP score?

Google rates INP of 200 milliseconds or less as good, more than 200 ms up to 500 ms as needs improvement, and more than 500 ms as poor. The value is judged at the 75th percentile of real page loads, separately for mobile and desktop. INP replaced First Input Delay as a Core Web Vital on March 12, 2024.

Do Core Web Vitals affect Google rankings?

Google's Search Central documentation says Core Web Vitals are used by its ranking systems, but also that there is no single page experience signal and that Search always tries to show the most relevant content even when page experience is sub-par. Google adds that chasing a perfect score just for SEO may not be the best use of your time.

Why does my PageSpeed Insights score differ from the Core Web Vitals assessment?

The 0 to 100 performance score comes from Lighthouse lab data collected in a simulated environment, while the Core Web Vitals assessment uses CrUX field data from real Chrome users over the previous 28 days. They measure different things, so a page can score well in the lab and still fail in the field, or the other way round.

  • #Core Web Vitals
  • #LCP
  • #INP
  • #CLS
  • #PageSpeed Insights
  • #Search Console
  • #technical SEO

t-trends

SaaS and AI tools, cross-border shopping, international money transfers and Web3 — compared using official pricing, terms and statistics.

About · Editorial Policy

↑