Conclusion first: this is not a coat of paint. Version 16.0 moves the anonymous browser off the older Firefox ESR 140 track and onto Firefox ESR 153, the long-support branch Mozilla now maintains, and Mozilla's public release-train tracker (whattrainisitnow.com) shows the 140 branch being retired in October 2026. Read plainly: the 15.x builds many people are still running now sit one engine generation behind on upstream security fixes. Everything else in the release is narrower in scope, and for the average reader the browsing experience is essentially unchanged — existing preferences are preserved during the transition, according to coverage of the announcement. So the honest verdict from the IT Insights general desk is "upgrade this week, expect nothing dramatic, and switch on one new option if you share your screen."
Where the update genuinely pays off is in three specific situations: you run Linux on Arm hardware, you are stuck on an old 32-bit machine, or your main device is an Android phone. Below we compare the two generations, split the outcome by platform, and finish with a checklist and the questions readers keep sending us.
| Decision-relevant dimension | Tor Browser 15.x (previous stable line) | Tor Browser 16.0 |
|---|---|---|
| Engine base | Firefox ESR 140 track | Firefox ESR 153, the currently maintained branch |
| Upstream security fixes | Backported only; the base branch is now past its active window | Full year of Firefox fixes carried into the stable release |
| Desktop settings | Legacy layout, switches spread across pages | Redesigned, grouped, with an updated offline manual |
| Window-title exposure | Real page title always visible to other apps | Optional "generic window titles" in Privacy and Security protections |
| Linux on Arm | Unsupported; the aarch64 gap was open through the 15.0 cycle | Native 64-bit Arm builds shipped |
| Linux 32-bit x86 | Still installable | Support ended; no further updates for those installs |
| Android interface | Older visual language | Updates aligned with Google's Material Design 3 |
| Upgrade friction | — | Low: existing preferences carry over |
Practically nothing changes in your routine, which is exactly the point: the value is under the hood. Tor support documentation recommends, separately from this release, that Windows users be on Windows 10 or 11 and Mac users on macOS 10.15 (Catalina) or newer (support.torproject.org) — so if you are still running the browser on older system software, the engine jump is the moment to fix that too, because those installations miss patches regardless of which browser version they carry.
This is the clearest gain. Community maintainers and users had been asking for months how to run the browser on aarch64 systems, with the answer through the 15.x cycle being effectively "not supported" (discourse.nixos.org). Tech coverage of the launch lists Raspberry Pi-class boards and Apple Silicon Macs running Linux as the obvious beneficiaries (9to5linux.com, linuxiac.com). If you previously kept a spare x86 machine switched on purely to reach Tor, that workaround is now obsolete.
Treat this as a hard deadline, not a preference. Dropping that architecture mirrors what Firefox itself did, and an installation that no longer receives updates is an installation accumulating known vulnerabilities while still being used for sensitive browsing. The proportionate response is to move to a 64-bit system, or to a device that can run one.
The interface work follows Google's Material Design 3 language, and the changelog frames these as interface updates rather than anything that changes how your traffic is routed. If the older menus felt cramped or inconsistent with the rest of the system, this is a quality-of-life improvement; if you only open the browser occasionally, you will barely notice it.
The feature exists because other applications on your machine can read window titles. The Tor Project's own notes on the release point at the everyday surfaces where that leaks: alt-tab and task switcher lists, screen shares, and the thumbnails meeting software draws from open windows (blog.torproject.org, x.com/torproject). Turning it on means those surfaces stop advertising which page you opened — valuable in a shared office, on a call, or in a living room with a second screen.
The cost is navigation: with titles genericised, several windows look identical at a glance, so keyboard switchers and anyone who juggles multiple sessions lose a shortcut. Our recommendation is situational rather than blanket — enable it when you present your screen or work in company, and disable it if you routinely manage several windows and rely on titles to tell them apart. It lives inside the Privacy and Security protections area of the redesigned Settings, and it is opt-in, so nobody loses anything by default.
Not reckless, but increasingly exposed: its engine line is the one Mozilla has rolled off, so the gap to current upstream fixes widens with each cycle. The low-cost answer is to move now.
Yes for this transition — preferences are preserved, and the update can be applied in place.
Only if you were running Linux on Arm without an official build. Boards and laptops on 64-bit Arm are now covered; anything that only boots a 32-bit x86 userland is the exception, and it is not covered going forward.
It reduces information available to local applications and to whatever is looking at your screen, titles included. Read it as a local-exposure control, not a network-level anonymity tool.
Desktop, on substance (engine, Settings structure, new privacy option, Arm builds). Android got the visible polish.
Measured against what it replaces, 16.0 is the rare engine swap that adds real capability without taking anything away from existing users: a supported path for Arm hardware, one opt-in protection that closes a small but persistent leak, a tidier control panel, and a cleaner interface on Android. The trade-off worth remembering is architectural rather than cosmetic — the release closes the door on 32-bit systems while opening one on 64-bit Arm. If you have been meaning to update for months, this is the release that makes the delay the more expensive option.