We opened the same 20 tabs in six browsers, five times each, and measured real RAM usage — including the browser we make.
Chrome has a reputation for eating RAM. So does Brave, in the opposite direction, for being the "efficient" pick. We stopped going by reputation and opened the same 20 tabs in six browsers, five times each, to measure what happened.
The browser that used the least RAM beat the one that used the most by 22%. That's not a rounding difference. It's the gap between one browser handling a heavy tab load comfortably and another one pushing your Mac toward a swap.
We opened the same 20 real websites in six browsers: Sigma, Chrome, Safari, Brave, Firefox, and Opera. Each browser ran the same test five times, with a full restart between runs, and we used the median of the five as the result, so a single unusual run couldn't skew the comparison.
Every browser used a clean profile with no extensions installed and no accounts logged in. Pages were opened in the same order in each browser, then given the same fixed wait time after the last tab loaded before we measured anything, so no browser got extra settling time that another didn't.
Memory was measured using macOS's physical footprint accounting, the same number Activity Monitor's Memory column shows, summed across every process belonging to the browser: the main process, per-tab renderer processes, GPU processes, and other helper processes. For Chromium-based browsers (Chrome, Brave, Opera, and Sigma), that means the main process plus every Helper and Renderer process. For Safari, it means Safari plus Safari Web Content, Safari Graphics and Media, and Safari Networking. For Firefox, it means the main process plus its GPU helper and Web Content processes. Measuring only the main process, which some comparisons do, misses most of what a browser uses once real tabs are open.
Test environment: a Mac with an Apple M3 chip, 16GB of unified memory, running macOS Tahoe 26.6.2. All six browsers were running their latest stable release as of September 11, 2026, the same day the test was conducted.
The 20 pages: Google Search, Wikipedia, MDN, GitHub, Hacker News, Stack Overflow, Apple, The Verge, CNN, BBC News, Reddit, YouTube, Google Maps, Canva, Figma, Notion, Spotify Web, Twitch, OpenAI, and Weather.com. YouTube and Twitch pages were left paused; no video was played, and no page was scrolled or interacted with during the test.
Sigma finished first at 6.5 GB, compared with 8.25 GB for Chrome and 8.30 GB for Safari, the two heaviest browsers in the test.
Compared with the winner, the other five browsers used:
Every one of the five runs, not just the medians:
All values in GB.
The medians tell part of the story; the spread tells the rest. Sigma's five runs stayed within 0.53 GB of each other, Firefox and Safari were similarly tight, and Opera settled into a narrow band after one higher first run. Chrome was the outlier here, swinging from 7.17 GB to 9.47 GB across its own five runs, a 2.3 GB spread on the exact same 20 tabs. That kind of inconsistency is itself a data point: whatever Chrome is doing with memory under this load, it isn't doing it the same way twice.
Sigma, in this test, by a clear margin over four of the five other browsers and a much narrower one over Brave specifically.
The genuine surprise here is Brave. Its reputation among people looking for a lightweight browser is strong, and the numbers back it up: 6.76 GB, just 4% behind Sigma and comfortably ahead of Firefox, Opera, Chrome, and Safari. If your priority is a fast, memory-conscious browser and switching away from Sigma isn't on the table for other reasons, Brave is the real second choice here, not an afterthought.
Chrome and Safari landed at the bottom of this specific test, both over 8 GB, both roughly a fifth to a fifth-and-a-half heavier than Sigma. That's worth sitting with for a moment: Safari is the browser built into every Mac, optimized by Apple for Apple hardware, and it still used more RAM than four of the five alternatives tested on that same hardware.
Chrome's process model is the direct answer, and it's by design, not a bug. Chrome runs each tab, and often each site within a tab, in its own separate process, a security feature called site isolation that keeps one compromised or crashing tab from taking down the rest of the browser or reading another site's data. The tradeoff Google itself has been open about is memory: more processes means more overhead, since each one carries its own baseline cost on top of whatever the page itself needs.
That architecture explains the ceiling, not the inconsistency we measured. The 2.3 GB spread across Chrome's five identical runs suggests something beyond the process model itself, possibly background tasks, extensions cleanup, or scheduling that didn't behave the same way twice. We can't isolate the exact cause from this test alone, but the instability is real and reproducible in our data, not a one-off.
The practical gap here isn't abstract. 1.75 GB is the difference between Sigma and Chrome at 20 tabs in this test, and that's memory the rest of your system, other apps, background processes, doesn't get to use. This test itself ran on a 16GB Mac, and the two heaviest browsers, Chrome and Safari, were already using roughly half of that total just to hold 20 tabs open. On a machine with that amount of memory, that gap is the difference between everything staying responsive and your system starting to swap to disk, which is where a computer starts to feel slow.
None of this makes Chrome or Safari bad browsers. It means that if keeping 15 or 20 tabs open is a normal part of how you work, the browser you pick has a measurable, repeatable effect on how much room everything else has to breathe.
Download Sigma to see how it handles your own tab habits, or compare it directly against Chrome in more depth in our full Sigma vs. Chrome comparison.
