Speed & performance · Explainer
PageSpeed Insights vs GTmetrix: Which Numbers to Trust
Short answer
Trust PageSpeed Insights field data for what real visitors experience, because it comes from Chrome users over 28 days. Trust GTmetrix for repeatable lab tests from a chosen location, a request waterfall and scheduled monitoring. The two tools share the Lighthouse engine, so their lab numbers differ mainly because of test location, device and connection settings.
Key takeaways
- PageSpeed Insights reports real-user field data from the Chrome User Experience Report over 28 days.
- Both tools run Lighthouse lab tests, so a different score often reflects different test settings, not a different page.
- A page passes Core Web Vitals only when LCP, INP and CLS are all good at the 75th percentile.
- The GTmetrix Grade blends a 60 percent Performance Score with a 40 percent Structure Score.
- PageSpeed Insights is free, while GTmetrix PRO starts at 5.99 dollars a month as of November 2026.
- Use PSI to judge visitor experience and GTmetrix to trace slow requests and watch a page over time.
What is the core difference between PageSpeed Insights and GTmetrix?
PageSpeed Insights shows real-user data plus one lab test, while GTmetrix focuses on repeatable lab tests with more control over location, device and connection.
That one sentence is the whole split. PageSpeed Insights reports field data from the Chrome User Experience Report (CrUX) and lab data from Lighthouse for one public URL. GTmetrix also runs Lighthouse, then adds its own Structure Score, a waterfall chart, saved history and scheduled monitoring.
Both tools can be right at the same time. A page can pass its real-user Core Web Vitals check in PSI and still score lower in a GTmetrix lab test run from a faraway server on a slow connection.
What does PageSpeed Insights actually measure?
PageSpeed Insights measures real-user experience over 28 days and one simulated Lighthouse load for mobile and desktop.
The top half is field data. Google's PSI page says it reports real users' First Contentful Paint (FCP), Largest Contentful Paint (LCP), Cumulative Layout Shift (CLS) and Interaction to Next Paint (INP) over the previous 28-day collection period. It also reports the experimental metric Time to First Byte (TTFB).
The bottom half is lab data. Lighthouse loads the page once in a simulated setting and gives a Performance score from 0 to 100, plus audits for accessibility, best practices and SEO. Google says a lab score of 90 or above is good, 50 to 89 needs improvement, and below 50 is poor.
Field data needs enough real traffic. Google says a new page or a low-traffic page may have no page-level data, so PSI falls back to origin-level data for the whole site. If the origin also lacks enough samples, PSI shows no field data at all.
What does GTmetrix actually measure?
GTmetrix measures one page load in a real browser with Lighthouse, then grades how fast it ran and how well it was built.
Its headline is the GTmetrix Grade, a letter from A to F. GTmetrix's report guide says the grade is a weighted average of a Performance Score at 60 percent and a Structure Score at 40 percent. A grade of 90 to 100 percent is an A, and 0 to 49 percent is an F.
The Performance Score comes from Lighthouse metrics such as LCP, Total Blocking Time and CLS. The Structure Score is GTmetrix's own check of Lighthouse opportunity and diagnostic audits, so it shows how well the page was built for speed. GTmetrix says Structure audits do not directly change the Performance Score.
The report also shows a waterfall chart, which lists every request in order with its size and time. That chart is the main reason many WordPress owners open GTmetrix after a PSI test.
Why do the two tools give different scores for the same page?
The scores differ because the test settings differ, even though both tools use Lighthouse.
PSI simulates a mid-tier Moto G4 phone on a mobile network for its mobile test, and an emulated desktop with a wired connection for desktop. Google says PSI runs in a Google data centre in North America, Europe or Asia, and the exact location can vary.
GTmetrix lets you pick the test location, connection speed and, on paid plans, the device.
Its PRO guide, updated November 18, 2026, says PRO users can test on more than 55 simulated mobile devices and monitor from up to 33 global test locations. A test from a server near your visitors will usually look faster than one from far away.
Lab scores also move from run to run. Google's PSI page lists local network availability, client hardware availability and client resource contention as common sources of variation. Run each tool 3 times and compare the middle result, not a single best or worst run.
Which numbers should you trust for SEO and real visitors?
Trust the PSI field data for visitors, and treat lab scores as a guide to what to fix.
Google's PSI page says field data captures true, real-world user experience, while lab data is useful for debugging in a controlled setting but may not capture real-world bottlenecks. That matches how the Core Web Vitals assessment works: a page passes only when the 75th percentile values for LCP, INP and CLS are all good.
The good thresholds, from web.dev, are LCP of 2.5 seconds or less, INP of 200 milliseconds or less, and CLS of 0.1 or less.
The poor thresholds are LCP over 4 seconds, INP over 500 milliseconds and CLS over 0.25. Each metric is judged at the 75th percentile, so at least 3 in 4 visits must meet the good mark.
That is why a lab score of 100 does not prove visitors are happy.
A site can score well in a single simulated load and still fail in the field if real visitors use slower phones or slower networks. For help acting on a failing field result, see how to pass Core Web Vitals in WordPress.
How do the scores and grades compare side by side?
The table below puts the two tools on the same footing so you can compare like with like.
| Feature | PageSpeed Insights | GTmetrix |
|---|---|---|
| Cost | Free, no account needed | Free Basic account; PRO from 5.99 dollars a month as of November 2026 |
| Real-user field data | Yes, CrUX over 28 days | No, lab data only in its main report |
| Lab engine | Lighthouse | Lighthouse, plus GTmetrix Structure audits |
| Headline number | 0 to 100 Performance score | GTmetrix Grade A to F, from 60 percent Performance and 40 percent Structure |
| Good lab score | 90 or above | Grade A, 90 to 100 percent |
| Pick test location | No, Google data centre varies | Yes, up to 33 locations on higher PRO plans |
| Pick device and connection | No, fixed Moto G4 mobile and emulated desktop | Yes, 55 plus simulated devices on PRO |
| Waterfall chart | No | Yes |
| Scheduled monitoring and alerts | No | Yes, hourly on select PRO plans |
| Report history | No saved history | Yes, 1 to 12 months by plan, up to 24 months on Enterprise |
The Lighthouse weights behind both lab scores are the same family of metrics.
Chrome's Lighthouse 10 post says CLS rose to 25 percent of the Performance score when Time to Interactive was removed. The other weights were Total Blocking Time at 30 percent, LCP at 25 percent, First Contentful Paint at 10 percent and Speed Index at 10 percent.
PSI moved to Lighthouse 13 on October 20, 2025, according to Google's PSI release notes. That update also adjusted CPU throttling for mobile lab tests, which Google said should raise mobile Total Blocking Time scores for many sites.
How much does each tool cost?
PageSpeed Insights is free, and GTmetrix is free to start with paid PRO plans for more tests, locations and history.
PSI needs no account and no payment. You paste a public URL at pagespeed.web.dev and get a report for mobile and desktop. That makes it the cheapest way to check field data for any page, including a rival page.
GTmetrix PRO starts at 5.99 dollars a month as of November 2026, according to GTmetrix's own alternatives page.
Its pricing page lists plans by on-demand tests, monitored slots and locations. The entry plan allows 50 on-demand tests a month, 1 monitored slot and 5 test locations, while higher plans allow 150 tests or unlimited tests, 3 to 50 monitored slots and 10 to 33 locations.
Retention also rises with the plan. GTmetrix's PRO guide says Lite, Lite Plus and Micro plans keep reports for up to 1 month, and Core and similar plans keep them for up to 6 months.
Expert, Champion and Team plans keep them for up to 12 months. GTmetrix also offers a 14-day money-back guarantee on PRO plans.
When should you pick PageSpeed Insights, and when GTmetrix?
Pick PageSpeed Insights to judge visitor experience, and pick GTmetrix to repeat a test, trace a slow request or watch a page over time.
Choose PSI when you need to answer one of these questions:
- Did real visitors pass Core Web Vitals in the last 28 days?
- Which metric, LCP, INP or CLS, is failing at the 75th percentile?
- Is the mobile result different from the desktop result?
Choose GTmetrix when you need to answer one of these questions:
- Which single request is slow, large or blocking the page?
- Does the page load slower from London than from Seattle?
- Did the score change after a plugin update, and can you prove it from saved history?
- Do you need an alert when a key page slows down?
For a WordPress site, the two tools fit different jobs in the same week. Use PSI to set the target from field data, then use GTmetrix to find the slow request and re-test after each fix.
If you want a sense of what a good load time looks like first, read how fast a WordPress site should load.
How do you use both tools together on a WordPress site?
Run PSI first for the verdict, then run GTmetrix with matching settings to find the cause.
- Paste the same public URL into PSI and note the field assessment, the 75th percentile LCP, INP and CLS, and the mobile and desktop lab scores.
- Open GTmetrix, sign in, pick the test location nearest to most of your visitors, and run the same URL 3 times.
- Open the waterfall chart and sort by time or size to find the slowest and largest requests.
- Fix one item at a time, such as a large hero image, a slow third-party script or a chat widget, then re-test in GTmetrix.
- Wait for field data to catch up before you judge the fix in PSI, because CrUX covers the previous 28 days.
Step 5 matters most. A fix made today will show in a new lab test today, but the PSI field section will still include weeks of older visits. That lag is normal, not a sign that the fix failed.
Common WordPress fixes sit in plain text here because each has its own guide: caching, image sizing, script delay and hosting choice.
Start with how to speed up WordPress for the order of fixes.
Check does hosting affect WordPress speed before you blame a plugin for a slow server response.
If search traffic is your main worry, read does page speed affect SEO to see where speed fits among ranking signals.
What mistakes should you avoid?
Avoid chasing a single lab number and treating it as a ranking score.
- Do not compare a PSI mobile score with a GTmetrix desktop score and call the gap a problem.
- Do not judge a fix from one run, because lab scores vary with network and server load.
- Do not trust a lab score of 100 as proof that visitors pass Core Web Vitals, because only field data shows that.
- Do not test a staging URL in PSI and expect field data, because PSI needs a public, crawlable URL with enough real samples.
- Do not pay for monitoring before you know which pages matter, because each URL and setting uses a monitored slot in GTmetrix.
- Do not fix every audit in order, because some audits save milliseconds while one slow request costs seconds.
The verdict by situation is simple. For a quick check of real visitors, PSI alone is enough. For a one-off fix on a slow page, run PSI once and GTmetrix 3 times.
For a shop or a site where speed affects sales, keep PSI as the monthly verdict and GTmetrix monitoring as the early warning.
Frequently asked questions
Can GTmetrix replace PageSpeed Insights?
No, because GTmetrix does not report real-user field data from Chrome users. It is a lab tool that runs Lighthouse from a chosen location and adds a waterfall, history and monitoring. Use it to find the cause of a slow load, and use PageSpeed Insights to judge what real visitors experienced over the last 28 days.
Why is my GTmetrix score higher than my PageSpeed Insights score?
The settings differ. PageSpeed Insights simulates a mid-tier Moto G4 phone on a mobile network, while GTmetrix may test on a desktop view with an unthrottled connection from a nearby server. Location, device, connection and server load all move a lab score, so compare runs only when the settings match.
Is a GTmetrix Grade A the same as a PageSpeed Insights score of 100?
No. A GTmetrix Grade A means the weighted grade reached 90 to 100 percent, blending 60 percent Performance Score with 40 percent Structure Score. A PageSpeed Insights score of 100 is a Lighthouse lab score only. Neither number proves that real visitors pass Core Web Vitals, which depends on field data.
Do I need a paid GTmetrix plan to test my WordPress site?
No. A free GTmetrix account lets you run lab tests and monitor one URL, with fewer locations and shorter report retention. Pay only if you need more on-demand tests, more monitored pages, more test locations, mobile device testing, hourly monitoring or longer history for client reporting.
How long does it take for a speed fix to show in PageSpeed Insights?
A lab test shows the change at once, because it loads the page again. Field data takes longer, because it covers the previous 28 days of real visits. Wait for enough new visits to enter that window before you judge whether the fix moved LCP, INP or CLS for real visitors.
Which tool should I use first after changing a WordPress plugin?
Use GTmetrix first, because you can repeat the same test from the same location and compare the waterfall before and after the change. Then check PageSpeed Insights later for the field result. That order gives you a fast lab signal now and a real-visitor verdict once new data arrives.
Sources
- About PageSpeed Insights Google for Developers
- How the Core Web Vitals metrics thresholds were defined web.dev
- Everything You Need to Know About GTmetrix Reports GTmetrix
- GTmetrix Pricing GTmetrix
- Getting Started With GTmetrix PRO GTmetrix
- What is new in Lighthouse 10 Chrome for Developers
- Why lab and field data can be different web.dev
Is your product missing, or something out of date?
Tell us what to add or correct. We read every note and reply when we need a source. A message does not buy a ranking.
Thanks — we review requests within 3 business days.
