Latest Chrome Release Fixes 213 Vulnerabilities, 7 Critical; 14 Reports Aided by AI

Google on November 3 released "Chrome 157," the stable desktop version of its Chrome web browser. This update includes 213 security fixes, with 7 rated at the highest severity level, "Critical." According to the official blog, the versions for Windows and macOS are 157.0.8103.38/.39 and the Linux version is 157.0.8103.38, with rollout proceeding gradually over the coming days to weeks. An Android version with the same fixes, 157.0.8103.38, was released the same day and will be made available through Google Play over several days.

The severity breakdown is 7 Critical, 77 High, 75 Medium, and 54 Low. The 7 Critical flaws are as follows.

Note: CVE numbers, affected components, and vulnerability types are based on Google's official announcement.

What draws attention in this round of fixes is the breakdown of reporters. Among the 7 Critical flaws, the use-after-free in Prerender (CVE-2026-121035) was reported by Anthropic's Xinyang Ge with the assistance of the company's AI, "Claude." In the list, reports explicitly credited as Anthropic-affiliated and Claude-assisted total 14 overall. In addition, 7 reports were

Looking at the bounty status, amounts were explicitly stated for only 9 reports, while 39 are listed as "TBD" and 165 as "N/A" (not applicable or undisclosed). The official announcement contains no indication that any of these vulnerabilities have already been exploited in attacks.

Just a week earlier, on October 27, Google had also fixed 52 vulnerabilities, including 4 Critical ones, in the Chrome 156 series update. Large-scale fixes are continuing in a short period. The desktop version can also be updated manually from the settings screen, and Google is urging users to update promptly.

Chrome is one of the most widely used browsers in the world across desktop and mobile, and the scale of fixes directly affects security management for both enterprises and individuals. Particularly in updates like this one that include multiple Critical issues, there are many situations where users are expected to apply fixes manually rather than waiting for automatic distribution. As AI-driven vulnerability discovery pushes up the number of reports, the speed of vendors' fix response and distribution is likely to be scrutinized even more closely going forward.