Can A Downloadable AI Turn Firefox Fixes Into Working Attacks?
AIThis post was created with the assistance of artificial intelligence (AI).

🔍 Read the full analysis: Can A Downloadable AI Turn Firefox Fixes Into Working Attacks? on ThorstenMeyerAI.com

Before you orderOffer from Amazon

Get hardware and tech essentials delivered free with Prime

  • Fast, free delivery on millions of items
  • Prime Video, Amazon Music and more included
  • Member-only deals all year
Start your free Prime trial Free trial for eligible customers · Cancel anytime
As an affiliate, we earn on qualifying purchases.

TL;DR

Epoch AI reports that the downloadable, open-weight model GLM 5.3 built working file-reading attacks against six Firefox releases in a controlled test, using security-fix information that was already public. Mozilla patched an additional, previously unpublished flaw used in the tests in Firefox 159; the researchers have not yet tested whether similar attacks work against that release.

Epoch AI researchers report that the downloadable open-weight model GLM 5.3 built attacks against six consecutive Firefox releases in a controlled test, turning public security-fix information into a webpage that could read a local file when opened. Mozilla released Firefox 159 on October 27, patching a separate, previously unpublished flaw the attacks used; Epoch has not yet tested whether similar attacks work against that release.

In a report published October 29, six Epoch AI researchers described giving the model a specific task: create a webpage that could read a file from a computer running a specified Firefox version. Their test used Ubuntu 26.04, default settings and official Firefox releases. The researchers said the model had no internet access, and the experiment was conducted in a lab rather than against real users.

Epoch said GLM 5.3 succeeded against all six releases from 156.0 through 158.0.1, issued between September 17 and October 26. Researchers allowed up to 48 hours for each attempt; the fastest successful run took 12 hours. After building the testing setup, Epoch estimated API fees at about $1,000 per version. It said developing and testing that setup cost more than $100,000.

The model used publicly visible Firefox changes to identify fixes that had not yet reached the stable release. Epoch said each successful attack on recent versions also used an unpublished flaw that the model rediscovered. The researchers reported that flaw to Mozilla, which patched it in Firefox 159. Epoch describes the result as a lab demonstration, not evidence that real users were attacked.

At a glance
reportWhen: Study published October 29, 2026; Mozil…
The developmentEpoch AI published a study reporting that GLM 5.3 turned public Firefox security fixes into working attacks against six releases in a lab test.
Crypto market snapshot
Fear & Greed Index
30/100 — Fear
Bitcoin BTC$76,913▼ 4.7%
Ethereum ETH$2,235▼ 4.3%
Tether USDT$0.9992▲ 0.0%
BNB BNB$699.3▼ 6.0%
XRP XRP$1.24▼ 3.9%
USDC USDC$0.9997▲ 0.0%
Solana SOL$94.18▼ 6.5%
TRON TRX$0.3328▲ 0.5%
Live data · CoinGecko · alternative.me (24h change)
Thorsten Meyer AI
Lab test · Epoch AI

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.

6 / 6Firefox releases compromised in Epoch’s test (156.0 to 158.0.1)
12 hfastest successful run; up to 48 h were allowed
≈ $1,000marginal API cost per Firefox version, once the setup existed
0refusals or safeguards triggered, at any provider

Where the attack happens

The life of a security fix, simplified

  1. 1 · Fix writtenDevelopers find and fix a security bug
  2. 2 · Fix visibleThe change shows up in public code and test builds, disguised as an ordinary change. The patch gap opens
  3. 3 · Update shipsThe fixed version is released to everyone
  4. 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

Firefox 156.0Firefox 156.0.1Firefox 157.0Firefox 157.0.1Firefox 158.0Firefox 158.0.1

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

Key flaw fixedNot yet re-tested by Epoch

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

Why Public Fixes Can Help Attackers

The finding highlights a patch gap: the period after a security fix becomes visible but before users install an update. In open-source software, code changes can reveal what was repaired. Epoch’s test suggests an AI model can help analyze those clues and build an attack more quickly than a manual process might allow.

That matters to users and software teams because a published fix can have two audiences: defenders applying it and attackers studying it. The study does not establish how often this method would work outside Epoch’s setup, or whether it can be used reliably against current browsers. It does show why installing security updates promptly remains relevant, while giving organizations reason to review how quickly they deploy patches.

Epoch also reported that its setup failed against Chrome after about $100,000 in attempts. That result is specific to the researchers’ test and should not be read as proof that Chrome is generally immune to AI-assisted vulnerability research.

Products Worth Considering

Apple MacBook Pro 2026 14.2 Laptop with M5 Pro chip, 15‑core CPU, 16‑core GPU, 24 GB Unified Memory, 1 TB SSD, Wi‑Fi 7 – Silver

Apple MacBook Pro 2026 14.2 Laptop with M5 Pro chip, 15‑core CPU, 16‑core GPU, 24 GB Unified Memory, 1 TB SSD, Wi‑Fi 7 – Silver

Apple MacBook Air 13.6-inch Laptop with M5 Chip, 24GB Memory, 1TB SSD, Midnight

Apple MacBook Air 13.6-inch Laptop with M5 Chip, 24GB Memory, 1TB SSD, Midnight

Apple MacBook Air 15-inch Laptop with M5 Chip, 16GB Memory, 512GB SSD, Liquid Retina Display, and Touch ID (Midnight)

Apple MacBook Air 15-inch Laptop with M5 Chip, 16GB Memory, 512GB SSD, Liquid Retina Display, and Touch ID (Midnight)

As an affiliate, we earn on qualifying purchases.

How the Firefox Patch Gap Works

Software developers often publish a security fix before every user has installed it. A change in public code or a release note can point to the part of a program that was vulnerable. The interval between that information becoming available and users receiving or applying the update is commonly called the patch gap.

Epoch asked GLM 5.3 to examine fixes not yet included in the stable Firefox version and allowed it to work with basic tools without further human input. According to the researchers, it selected bugs, tested approaches and adjusted after crashes. The study says the task was framed as security research, but Epoch found no provider safeguards or refusals during its attempts; that is the researchers’ account, not an independent audit of all model safeguards.

Mozilla’s Firefox 159 security advisory lists 167 fixes, according to the supplied report: 108 rated high, 52 moderate and seven low. The source notes that Mozilla had changed its counting approach, so the total should not be treated as a direct measure of a sudden increase in risk.

“widely available open-weight models can now do much, or even all, of this work on their own.”

— Epoch AI researchers

What Firefox 159 Testing Could Show

Epoch has not reported testing whether the same attack approach succeeds against Firefox 159. Its expectation that Mozilla’s fix will make attacks substantially more difficult is not a confirmed test result. The supplied material also does not establish whether attackers outside the lab knew of or used the flaw before it was patched.

The results cover six specified Firefox releases on a particular Ubuntu setup. Epoch said earlier Firefox versions were likely vulnerable but were not systematically tested. The study does not show that the model can take over a computer: the demonstrated objective was reading a local file. Nor does it establish that users were targeted or harmed.

Epoch reported no safeguards or refusals from the providers involved, including GLM 5.3’s maker, Z.ai. Further details about provider policies, the model’s performance under different conditions, and the extent to which other systems could reproduce the results remain unclear from the supplied account.

Updates and Follow-Up Tests

Firefox users should install version 159 or later and restart the browser, as the source report advises. The supplied material does not specify whether Mozilla has issued a later update, so users should check the browser’s update function for the current release rather than relying only on the version number in this report.

The next key research milestone is Epoch’s assessment of Firefox 159 and whether the fix changes the attack’s feasibility. Further testing could also clarify how often the approach works across different systems and browser configurations. Mozilla’s patch addresses the reported unpublished flaw, but the supplied information does not include a Mozilla statement on the study or a full technical account of its response.

For organizations, the immediate operational question is how quickly security updates can be reviewed and deployed. The experiment offers evidence that public fixes may be analyzed quickly in a controlled setting; it does not establish a universal deadline or prove that every patch can be turned into an attack.

Key Questions

What did GLM 5.3 do in the test?

Epoch AI says the model built a webpage that could read a local file when opened in one of six specified Firefox releases. The work took place in a controlled lab, not on real users’ computers.

Were real Firefox users attacked?

The supplied report describes a lab experiment and provides no evidence that real users were attacked or that files were stolen.

Did Mozilla fix the flaw used in the tests?

Epoch says it reported an unpublished flaw to Mozilla, which patched it in Firefox 159, released October 27, 2026. Epoch has not yet tested whether similar attacks work against that version.

What should Firefox users do?

Install the latest available Firefox security update and restart the browser. The source report specifically advises version 159 or later, while the current available release may have changed.

Does the study show that AI can attack any browser?

No. The results cover specified Firefox versions and a particular test environment. Epoch said its setup failed against Chrome after about $100,000 in attempts, but that single result does not establish broad immunity for Chrome or other browsers.

Source: ThorstenMeyerAI.com

Nothing in this article is financial or investment advice. Cryptocurrency and precious-metal investments carry significant risk — do your own research and consider a licensed advisor.
You May Also Like

Trade and supply-chain operations signal monitor: US-Iran talks to begin Sunday in Switzerland as Tehran closes the strait over Lebanon fi

U.S.-Iran negotiations are set to begin Sunday in Switzerland, with tensions rising as Tehran closes the Strait over Lebanon, impacting global trade routes.