WordPress 6.5 performance improvements

This post is the latest in a series of updates focused on the performance improvements of major releases (see 6.4, 6.3, and 6.2).

WordPress 6.5, “Regina” is the first major version of WordPress released in 2024. This release includes several important performance enhancements impacting the user experience for site visitors, along with remarkable improvements to editor performance. Importantly, WordPress 6.5 delivers site performance that is similar, if not superior to previous versions, despite the addition of many significant new features.

In our analysis of the Twenty Twenty-four theme, we observed modest changes in website front-end performance. The median LCP time shows a slight 0.81% decline in non-translated tests, with a 0.95% improvement in translated tests. Similarly, the Twenty Twenty-one theme exhibits a 1.13% drop in median LCP times in non-translated tests, yet showcases a notable 4.50% improvement in translated tests. Further details on the methodology behind these measurements are provided later in this article.

class="wp-block-heading">Key improvements

Improved performance for translated sites

This version of WordPress includes a new localization system that loads translation files more quickly, and introduces support for providing translation files as native PHP files. As the merge announcement summarizes, these enhancements are expected to bring a 23.5% improvement in loading time and 41.4% reduction in memory consumption for translations.

A significantly faster editing experience

A major focus of this release was improving performance while editing your site. WordPress 6.5 delivers 5x faster typing processing, 2x faster editor loading, and a 60% reduction in pattern loading, based on measurements collected as part of the overall effort documented in this GitHub issue.

Improvements for registering block variations

The WordPress block editing system allows blocks to be registered with a set of block variations, which makes it easy to define different versions of a block without needing to duplicate the whole block. This version of WordPress adds support for registering block variations only when used, avoiding costly processing when this data is unnecessary, which makes server rendering 5% faster.

Support for AVIF image format

WordPress 6.5 supports AVIF, a modern image format that offers significant improvements in image quality and compression over previous formats like JPEG, PNG, and even WebP.  AVIF images can be up to 50% smaller than JPEGs while maintaining the same image quality.

Additional performance focused changes

In total, there were 20 performance related improvements included in this release, split evenly between new enhancements (10) and bug fixes (10).

How release performance is measured

The performance measurements used for the overview are based on benchmarks1 conducted using an automated workflow on GitHub action runners. Benchmarks were taken of the homepage of the Twenty Twenty-one, Twenty Twenty-three, and Twenty Twenty-four themes with and without translations installed, comparing WordPress 6.5 with WordPress 6.4.3 (the latest version of WP 6.4 available when 6.5 was released).

Performance metrics were collected from 100 runs for both Core Web Vitals (CWV) and Server-Timing headers provided by the Performance Lab plugin using CLI scripts from the WPP Research repo.

Benchmark Data

No translation (en_EN):

Translation (it_IT):

Follow up from this release

Each release, the Performance Team looks for opportunities to improve the performance of WordPress for the following releases, which includes identifying ways we can improve the tooling and processes we use to support the performance practice. For example, we’re collecting opportunities to improve our performance testing in this GitHub issue

You can also follow progress on other performance-related work being planned for the WordPress 6.6 release in Trac and in the Gutenberg repository. A full list of the Performance Team’s priorities for the year is available on the 2024 Roadmap page in the team’s handbook. Come join us in making WordPress as performant as possible.

Props to @flixos90, @peterwilson, @adamsiverstein, @annezazu, and @jorbin for contributing to this post.

  1.  Benchmark measurements use lab data to gather performance metrics under controlled conditions and may not reflect the way the software performs in the field. For more on the differences between Lab and Field data, see this article.

#6-5, #core, #core-performance, #performance

Leave a Reply

Your email address will not be published. Required fields are marked *