🔍 Read the full analysis: Lessons From Epoch AI’s Demonstration Of AI And Firefox Attacks on ThorstenMeyerAI.com
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
Epoch AI reports that open-weight model GLM 5.3 built working attacks against six Firefox releases by analyzing public security fixes before those fixes reached stable users. Mozilla patched an additional, previously unpublished vulnerability used in the attacks in Firefox 159, but Epoch has not yet tested whether the approach works against that release.
Epoch AI reported that open-weight model GLM 5.3 built working attacks against six consecutive Firefox releases, using public information about security fixes before those fixes reached stable users. Mozilla has since patched a second vulnerability used in the attacks in Firefox 159, released October 27; the researchers have not yet tested whether their method works against that version.
In the controlled test, Epoch asked GLM 5.3 to create a webpage that could read a local file when opened in a specified Firefox release. The researchers say the model succeeded against versions 156.0 through 158.0.1, released between September 17 and October 26. The fastest successful run took 12 hours; Epoch allowed as long as 48 hours for each attempt.
The model used information already available in public code, including fixes that had not yet reached the stable browser. Epoch describes the interval between a fix becoming visible and users installing it as the patch gap. According to the report, each successful attack against recent releases also used a separate vulnerability that had not been publicly disclosed. Epoch reported that flaw to Mozilla, which addressed it in Firefox 159.
Epoch estimates that running an attack against each new release cost about $1,000 in API fees after the researchers had built their setup. The research team says developing and testing that setup cost more than $100,000. Those figures describe this experiment, not the cost of all AI-assisted vulnerability research or attacks.
Infographic 01 · The patch gap
A published fix is a map. GLM 5.3 read it.
How an open-weight model turned Firefox security fixes into a working attack before the update reached users, in Epoch AI’s offline lab test.
Where the attack happens
The life of a security fix, simplified
- 1 · Fix writtenDevelopers find and fix a security bug
- 2 · Fix visibleThe change shows up in public code and test builds, disguised as an ordinary change. The patch gap opens
- 3 · Update shipsThe fixed version is released to everyone
- 4 · You install and restartOnly now is your browser protected. The gap closes for you
GLM 5.3 worked in step 2: it picked out the security fixes and turned them into a working attack before step 3.
Compromised in Epoch’s test
Released 17 Sep to 26 Oct 2026 · only information public at release
Result: a web page that read a local file and sent it to the attacker’s server when visited. No further user action.
Firefox 159.0 · 27 Oct
Fixes the unpublished flaw every recent success relied on
Epoch expects the fix to make similar attacks “at least substantially more difficult”.
Sources: Epoch AI, 29 Oct 2026; Firefox 159.0 release notes. Lab test without internet access; no real users were attacked. No exploit details are shown here.
Powered by Thorsten Meyer AI
Public Patches Can Shorten Attack Timelines
The study draws attention to a familiar security problem: once a software fix is public, its details can help defenders understand a vulnerability, but can also give attackers clues about what changed. Epoch’s result suggests that a widely available model may be able to analyze those clues and assemble an attack within hours, rather than relying only on specialist researchers working manually.
For users, the immediate practical implication is to install Firefox 159 or later and restart the browser, as the source report advises. For software makers and organizations, the test highlights the importance of reducing the time between publishing a fix and getting it installed. The study does not show that ordinary users were attacked, or that a model can automatically compromise any browser or computer.
Products Worth Considering
Bose QuietComfort Wireless Bluetooth Headphones with Active Noise Cancellation, Mic, USB‑C Charging, Deep Bass, 24‑Hour Battery – White Smok
Amazon Kindle Paperwhite (12th Generation, 2024), 16 GB, 7-inch Display, Jade
Samsung 27 Odyssey OLED G5 (G50SF) QHD QD-OLED Gaming Monitor
As an affiliate, we earn on qualifying purchases.
How the Firefox Patch Gap Worked
Security updates often reveal information about a flaw through public source-code changes, test releases or advisories. That information can appear before every user has installed the stable update. In this experiment, Epoch gave the model a Firefox release and public information available at that time, then allowed it to work without further input. The researchers say it selected relevant fixes, tested approaches and adjusted after failures.
Epoch chose a local-file read as its target because access to browser-accessible files could expose sensitive material, such as private documents or login data. The experiment ran on a test machine using Ubuntu 26.04 and an official Firefox release. The researchers say it had no internet access, so the test could not target a real website or user. They also report that the attack chain required a second flaw to get past Firefox’s sandbox, the browser’s isolation boundary.
Epoch says the same setup did not succeed against Chrome after about $100,000 in attempts. That result is limited to the researchers’ test and should not be read as proof that Chrome is immune to other methods. The source also says earlier Firefox versions may have been vulnerable, but they were not systematically tested.
““Widely available open-weight models can now do much, or even all, of this work on their own.””
— Epoch AI researchers, as quoted in the report
Firefox 159 Has Not Been Tested
The report describes a controlled laboratory test, not an attack on real Firefox users. Epoch had not tested its setup against Firefox 159, so the effect of Mozilla’s fix on this specific attack chain remains unverified by the researchers. The source says the fix is expected to make similar attacks harder, but does not establish that every related risk has been removed.
Other limits also matter. The test covered six releases, not every browser version, operating system or configuration. Epoch says earlier releases are likely vulnerable but were not systematically examined. The Chrome result is likewise specific to this setup and does not establish that other attack approaches would fail. The source reports no evidence that the lab attacks were used against people outside the test environment.
Install Firefox 159 and Watch for Testing
Users following the report’s advice should update to Firefox 159 or later and restart the browser so the update takes effect. Mozilla’s Firefox 159 security advisory, identified in the source as MFSA 2026-118, lists 167 security fixes: 108 rated high, 52 moderate and seven low. The report cautions that Mozilla changed its counting approach, so those totals should be read in that reporting context.
The next evidence point is whether Epoch or other researchers test the attack method against Firefox 159 and publish results. The study also leaves open how well similar methods transfer to other software and how quickly fixes reach users in practice. Until further testing is reported, the confirmed finding is narrower: GLM 5.3 succeeded against six specified Firefox releases in Epoch’s lab, and Mozilla patched the additional flaw Epoch reported.
Key Questions
What did Epoch AI demonstrate?
Epoch says GLM 5.3 built working webpage-based attacks against six Firefox releases, using public information about security fixes that had not yet reached stable users.
Were real Firefox users attacked?
The report describes a lab test with no internet access and says the setup could not target real websites or people. It does not report attacks on real users.
What did Mozilla patch?
Epoch reported a separate, previously unpublished vulnerability used in successful attacks. Mozilla fixed it in Firefox 159, released October 27, 2026.
Does the study show Firefox 159 is vulnerable?
No. Epoch had not tested its setup against Firefox 159 when the report was written. The researchers expected the fix to make similar attacks substantially more difficult, but that expectation was not yet confirmed by a test.
What should Firefox users do?
Update to Firefox 159 or later and restart the browser, following the advice in the source report.
Source: ThorstenMeyerAI.com
