Lighthouse vs PageSpeed Insights: why scores differ (2026)
Lighthouse vs PageSpeed Insights explained: lab vs field data, why your scores differ between tools and runs, and which number Google actually uses.
Lighthouse vs PageSpeed Insights is one of the most common points of confusion in web performance: the two often show different numbers for the same page.
They're closely related. PageSpeed Insights runs Lighthouse on Google's servers and adds real-user data from Chrome on top. Lighthouse on its own is the testing engine, and you can run it in Chrome DevTools, from the command line or through tools built on the PageSpeed Insights API. Once you know which numbers come from a lab test and which come from real visitors, the differences make sense.
Lighthouse vs PageSpeed Insights at a glance
| Lighthouse | PageSpeed Insights | |
|---|---|---|
| What it is | An open-source auditing tool | A Google web service |
| Where it runs | Your browser, your machine or a server | Google's servers |
| Lab data | Yes | Yes, it runs Lighthouse |
| Real-user (field) data | No | Yes, from the Chrome UX Report when available |
| Categories | Performance, Accessibility, Best Practices, SEO | The same four, plus a Core Web Vitals assessment |
| Device profiles | Mobile or desktop, your choice | Separate mobile and desktop tabs |
Lab data vs field data
Lab data comes from one simulated visit. Lighthouse loads the page under fixed conditions, on mobile a simulated mid-range phone on a throttled connection, and measures what happens. It's repeatable and shows causes, so it's the right tool for debugging.
Field data comes from real Chrome users who visited the page over the previous 28 days, collected in the Chrome UX Report. It reflects real devices, networks and locations. PageSpeed Insights shows it at the top of the report when Google has enough traffic data for the page or site; for low-traffic pages it may not be available at all.
The two answer different questions. Lab data asks how this page performs under test conditions; field data asks how real visitors actually experience it.
Why your Lighthouse score differs between tools
The same page can score differently depending on where Lighthouse runs:
- Hardware and network. DevTools runs on your computer and connection; PageSpeed Insights runs on Google's servers. Throttling is simulated, but the underlying machine still affects the result.
- Browser extensions. Extensions can inject scripts into the page during a DevTools test. Use an Incognito window.
- Lighthouse version. Different tools can run different versions, and scoring changes between versions.
- Device profile. Mobile scores are usually much lower than desktop. Compare mobile with mobile.
- Location. The test server's distance from your host and CDN affects timings.
Why your score changes between runs
Even in the same tool, scores move a few points between runs. Third-party scripts, ads, A/B tests, server load and network variation all change from one load to the next. Treat a single run as an estimate: run important tests several times and compare the middle result, not the best one.
Core Web Vitals vs the Lighthouse score
Core Web Vitals are three field metrics Google uses as part of its page experience signals:
- Largest Contentful Paint (LCP): good at 2.5 seconds or less
- Interaction to Next Paint (INP): good at 200 milliseconds or less. INP replaced First Input Delay in March 2024
- Cumulative Layout Shift (CLS): good at 0.1 or less
A page passes when at least 75% of real visits meet each threshold. Lighthouse can measure LCP and CLS in the lab, but it can't measure INP, because a lab test has no real user clicking. Its Total Blocking Time metric is the closest lab stand-in for responsiveness.
Which score does Google use for rankings?
Google's page experience signals use Core Web Vitals from real users, not the Lighthouse score. A page can score 70 in a Lighthouse lab test and still pass Core Web Vitals in the field, or the other way round. In practice, use Lighthouse to find and fix problems, and field data to confirm that real visitors feel the difference.
Why is my PageSpeed Insights score lower than Lighthouse in DevTools?
Usually because of the device profile and the machine. PageSpeed Insights defaults to its mobile tab with throttling, while a DevTools run may be set to desktop and use a fast laptop. Compare the same profile, run each test a few times, and the gap usually narrows.
Is PageSpeed Insights the same as Lighthouse?
Not quite. The lab half of a PageSpeed Insights report is a Lighthouse run, so the four category scores come from Lighthouse. PageSpeed Insights adds the field data from real Chrome users and the Core Web Vitals assessment, which Lighthouse on its own can't provide. Think of Lighthouse as the engine and PageSpeed Insights as one of the dashboards built on it.
Why is my mobile Lighthouse score so much lower than desktop?
Because the mobile test is deliberately harder. It simulates a mid-range phone with a slower processor on a throttled mobile connection, while the desktop test assumes a fast computer and connection. Heavy JavaScript costs far more time on the simulated phone, which shows up in Total Blocking Time and LCP. Since most visitors to many sites are on phones, the mobile score is usually the one to work on.
Why does PageSpeed Insights show no field data for my page?
The Chrome UX Report only includes pages and sites with enough real Chrome traffic. Newer and low-traffic pages often have no field data yet. In that case, rely on lab data and check again as traffic grows.
How to read a PageSpeed Insights report
A PageSpeed Insights report has two halves, and it helps to read them in order:
- Discover what your real users are experiencing. This is the field data. It shows whether the page or the whole site passes the Core Web Vitals assessment, with LCP, INP and CLS at the 75th percentile, plus a few other metrics. If it says there isn't enough data, skip to the second half.
- Diagnose performance issues. This is the Lighthouse lab run: the four category scores, the five lab metrics behind the Performance score, and a list of opportunities and diagnostics with the estimated savings for each.
Read the field data to decide whether you have a real-user problem, then use the Lighthouse half to find the cause. If the field data passes but the lab score is orange, you may not have an urgent problem. If the field data fails, fix it even if the lab score looks fine.
A worked example (made-up numbers)
Imagine a shop's product page. Lighthouse in DevTools on a fast laptop, set to desktop, shows Performance 96. PageSpeed Insights on its mobile tab shows Performance 58, and its field data says LCP is 3.1 seconds at the 75th percentile, which fails.
Nothing is broken: the first number is a desktop lab run on fast hardware, the second is a throttled mobile lab run, and the field data reflects real phones. The mobile lab run and the field data agree that the page is slow on phones, so that's where to work. Shrinking the hero image and deferring a review widget would be sensible first fixes, measured on the same mobile profile in the Lighthouse checker.
Which tool should I use?
Use PageSpeed Insights or a checker built on it for a quick, shareable result with field data where available. Use Lighthouse in DevTools when you're debugging locally and want to test changes before you deploy. Use Search Console's Core Web Vitals report to track real-user performance across your whole site.
How FeedPulse measures Lighthouse scores
The FeedPulse Lighthouse checker calls Google's PageSpeed Insights API on the mobile profile and shows all four category scores: Performance, Accessibility, Best Practices and SEO. Results are cached, for 6 hours on sites running FeedPulse and up to 14 days for one-off lookups, and you can re-run a check from the result panel. If Google's API hasn't returned a result yet, the badge says Measuring… instead of showing a placeholder number.
Related free tools
- Free Lighthouse checker — four Lighthouse scores for any URL, with an auto-updating badge
- How to improve your Lighthouse score — what to fix first in each category
- Free page speed test — mobile load time from PageSpeed Insights
- Website trust score checker — uses the four Lighthouse scores for half its points