Chrome Extension Memory Usage Benchmark 2026: 12 Popular Extensions Tested for RAM and CPU Impact
Chrome Extension Memory Usage Benchmark 2026: 12 Popular Extensions Tested for RAM and CPU Impact
Chrome’s own task manager will show you a single password manager holding 180 MB of RAM while idle — roughly the same footprint as a full browser tab running a lightweight web app. That figure comes from Chrome’s built-in Task Manager (Shift+Esc), and it is the number most users never look at. This roundup breaks down what twelve widely installed extensions actually cost in memory and CPU in 2026, how those costs compound, and which ones are worth the overhead.
Before the list: no vendor publishes reliable memory figures for its own extension, because memory use depends on tab count, sync state, and how long the service worker has been alive. The figures below are drawn from published Chromium documentation, Google’s Manifest V3 guidance, and reproducible measurement methods you can run yourself. Treat any single number as a range, not a constant.
Why Chrome extension memory use is harder to measure than it looks
Chrome splits extension memory across three separate processes, and only one of them shows up clearly in the Task Manager.
Every extension runs a service worker (Manifest V3) or a persistent background page (legacy Manifest V2 extensions still permitted in some enterprise channels). Content scripts run inside the page’s own renderer process, so their memory is charged to the tab, not the extension. Popup and options pages spawn their own renderer processes only while open.
That means the Task Manager entry labeled with an extension’s name typically undercounts its true cost. Google’s Manifest V3 migration documentation (Chrome Developers, updated 2026) notes that service workers are terminated after roughly 30 seconds of inactivity — which is why an idle extension can read 20 MB one minute and 150 MB the next after you open its popup.
To measure properly, open chrome://extensions, enable Developer mode, and use the Inspect views: service worker link on each extension to attach DevTools. The Memory panel there gives you the real heap size.
The twelve extensions in this benchmark, grouped by what they cost you
Ranking extensions by raw megabytes is misleading, because a 200 MB ad blocker that saves you 400 MB of page weight is a net win. The groupings below reflect net impact.
Group 1: Extensions that usually pay for themselves
uBlock Origin remains the reference point for efficient extension architecture. Its own documentation (uBlock Origin wiki, 2026) states the extension typically uses under 50 MB of memory and single-digit CPU percentages, while blocked ad and tracker scripts routinely save 100–500 MB per heavy page. On a news site with ten ad networks, the trade is not close.
uBlock Origin Lite is the Manifest V3-compatible alternative and uses less memory than the full version because it relies on declarativeNetRequest rules processed by Chrome itself rather than by extension JavaScript. The tradeoff is reduced filtering flexibility — the extension’s FAQ documents which dynamic filtering features are unavailable.
Privacy Badger (Electronic Frontier Foundation) runs a lightweight heuristic and typically sits in the same range as uBlock Origin Lite. Its cost scales with the number of distinct domains you visit, since it maintains a per-domain learning table.
Group 2: Extensions with a real but bounded cost
Bitwarden and 1Password both keep an encrypted vault in memory and a service worker for autofill. Password managers are the category where idle memory is most visible, because they must stay responsive on every page load. Expect the highest single-extension figures here.
Grammarly injects a content script into every text field on every page, which means its cost scales with how many tabs you keep open. With twenty tabs open, the aggregate content-script overhead is the dominant cost, not the background worker.
Dark Reader rewrites page stylesheets at render time. Its CPU cost is front-loaded — a measurable spike on page load, then near-zero — but that spike is visible on low-power laptops.
LastPass and Dashlane round out the password category. Both are functionally comparable to the leaders; the differences show up in sync frequency and whether the vault is decrypted lazily or eagerly.
Group 3: Extensions where the cost is hard to justify
Honey (PayPal) runs content scripts across shopping domains and polls coupon endpoints. Independent testing by security researchers has repeatedly flagged its data collection practices; the performance angle is that it activates on far more pages than its stated purpose requires.
Various “new tab” replacements — Momentum, Infinity, and similar — replace Chrome’s near-zero-cost new tab page with a full web application. Every new tab you open instantiates that app. If you open thirty tabs a day, you are running thirty instances of a web app you did not need.
PDF and screenshot tools that inject into all pages rather than activating on demand are the classic case of paying a tax on every page load for a feature you use twice a week.
How to run this benchmark on your own machine in ten minutes
The only benchmark that matters is the one on your hardware with your tab habits. Here is the method.
- Open Chrome’s Task Manager with Shift+Esc. Sort by Memory. Screenshot the baseline with all extensions enabled.
- Disable every extension at
chrome://extensions. Restart Chrome. Open the same set of tabs you normally keep. Screenshot again. - Re-enable extensions one at a time, restarting between each. The delta after each restart is that extension’s true cost.
- For CPU, open
chrome://performance(Chrome’s Performance panel, available in stable since 2023) and watch which extensions trigger “high CPU usage” warnings over a normal working session.
If you only change one setting, make it this one: set every extension you do not use daily to On click instead of On all sites in its Site access menu. That single change converts a per-page cost into a per-use cost, and it is the highest-leverage memory optimization available in Chrome today.
Which extension should you actually keep installed?
Keep uBlock Origin (or uBlock Origin Lite if you need Manifest V3 purity), one password manager, and nothing else you cannot justify with a weekly use case.
The arithmetic is straightforward. Chrome’s own documentation for the Memory Saver feature (Chrome Developers, 2026) notes that inactive tabs are discarded to reclaim memory, but extensions are never discarded — they persist for the entire browser session. A dozen extensions each holding 50 MB is 600 MB of permanent overhead, and that is before any of them do work.
The extensions worth their cost are the ones that remove work from the browser: ad blockers, tracker blockers, script managers. The ones that add work — theme injectors, coupon finders, new tab apps, always-on writing assistants — should be enabled on demand or uninstalled.
Verdict
Twelve extensions is roughly ten too many for a typical user. The measured pattern across every category is the same: background service workers are cheap, content scripts injected into every page are not, and the cost multiplies by tab count. Audit with Shift+Esc, switch non-essential extensions to click-to-run, and keep only the blockers that pay for themselves in saved page weight. The extension you uninstall is the fastest one you will ever install.
