Lighthouse updates can change PageSpeed Insights lab scores even when the page itself has not been modified. Version 13.6.0, for example, adjusted the loading simulation, so score increases across versions cannot be attributed entirely to site optimization. Before claiming a performance improvement, check the Lighthouse versions in both reports and compare the results with field data from real users. If the versions differ, acknowledge that some of the score change may come from the testing method.
13.6.0 changes the simulation, not necessarily the page’s speed
According to a report dated November 2, 2026, PageSpeed Insights had adopted Lighthouse 13.6.0, released on October 27. Google’s Barry Pollard said the new version simulated First Contentful Paint (FCP) and Largest Contentful Paint (LCP) more accurately, generally resulting in higher scores. This describes the simulation results; it does not guarantee a score increase for every site.
Pollard also noted that JavaScript-heavy sites were more likely to be affected, but the change is not exclusive to single-page applications (SPAs). A site’s architecture alone therefore cannot tell you whether its score will rise. Nor can the size of a score increase be taken as a measure of how much the user experience has improved.
Check versions before deciding why scores changed
The report notes that each PageSpeed Insights report displays the Lighthouse version beside the simulated device. When documenting optimization results, first check:
- Whether both reports use the same version: Tests before and after the 13.6 update use different simulation methods, so their scores should not be treated as measurements on the same scale.
- Whether optimization coincided with a version change: If both happened at the same time, a higher score may reflect both page modifications and changes to the testing method, rather than page modifications alone.
- Whether the score trend jumps at the update: Even without page changes, a tracking chart may show a jump when the version changes.
For example, a page’s score might rise after optimization, but the earlier report used an older version and the later one used 13.6.0. A more cautious description would be: “The lab score improved, but the version change may account for part of the difference,” rather than simply declaring that the site is now faster.
Field data answers a different question
Lab data shows simulated loading results under fixed device and network conditions. The report’s field data comes from CrUX and reflects real-user visits over the past 28 days. Lab data lets you examine test results, while field data helps you assess real-world user experience. Neither can replace the other.
Checking that the versions match is therefore just one check to make before comparing results; it does not prove that performance improved. When consulting field data, remember that it covers the past 28 days. Do not present it as an immediate result of a particular change. A better lab score alone still does not prove that real users’ experience improved at the same time.
Lighthouse updates and PageSpeed scores: perspectives from across the web
Why can scores still vary between tests using the same Lighthouse version?
Even with the same Lighthouse version, repeated tests can produce different scores. Network conditions, browser extensions that alter network requests, and the ads displayed on the page during a particular test can all affect the results. Confirming that the versions match helps rule out version differences, but you should not assume that a score change between individual tests is entirely due to page modifications. Consider these testing conditions when comparing results, too.
What exactly changed in the loading simulation in 13.6.0?
According to reporting from November 2026, Lighthouse 13.6.0 modified Lantern, the simulation used to predict loading performance. It uses the render-blocking flags Chrome provides for individual requests, falling back to fetch priority values when those flags are unavailable. These are changes to the simulation method, so score changes alone cannot establish how much the page itself has changed.
References
- Google Lighthouse Update Changes How PageSpeed Insights Simulates Loadingwww.searchenginejournal.com · 2026-11-02T21:52:57+00:00 — Source used in the article above.
- Google Lighthouse: What It Is & How to Use Itwww.semrush.com · 2021-09-13T19:32:00Z — Discusses: Why can scores still vary between tests using the same Lighthouse version?
- Google Rank Data Issues, Lighthouse Update & AI Opt-Out – SEO Pulsewww.searchenginejournal.com · 2026-11-07T13:00:00+00:00 — Discusses: What exactly changed in the loading simulation in 13.6.0?


