{"id":1907,"date":"2026-07-21T18:41:17","date_gmt":"2026-07-21T18:41:17","guid":{"rendered":"https:\/\/thedigitalfortress.us\/?p=1907"},"modified":"2026-07-21T18:41:17","modified_gmt":"2026-07-21T18:41:17","slug":"google-launches-gemini-3-5-flash-cyber-ai-to-find-and-fix-software-vulnerabilities","status":"publish","type":"post","link":"https:\/\/thedigitalfortress.us\/?p=1907","title":{"rendered":"Google Launches Gemini 3.5 Flash Cyber AI to Find and Fix Software Vulnerabilities"},"content":{"rendered":"
\n

\ue804<\/i>Ravie Lakshmanan<\/span>\ue802<\/i>Jul 21, 2026<\/span><\/span>Software Security \/ Artificial Intelligence<\/span><\/p>\n<\/div>\n

\n
<\/a><\/div>\n

Google’s DeepMind on Tuesday announced the release of Gemini 3.5 Flash Cyber<\/strong>, a specialized artificial intelligence (AI) model built atop 3.5 Flash that’s designed to discover, validate, and patch vulnerabilities quickly and efficiently.<\/p>\n

According to the tech giant, the model will be exclusively available to governments and trusted partners via CodeMender as part of a limited-access pilot program. CodeMender is an AI-powered agent for vulnerability discovery and patching that was unveiled by the company in October 2025.<\/p>\n

A Google DeepMind spokesperson told The Hacker News that there are plans to extend the model’s capabilities to include red-teaming features and end-to-end enterprise defense.<\/p>\n

The lightweight model, per DeepMind, is both cost-efficient and highly capable alternative to large, costly cybersecurity-focused models. CodeMender can call upon 3.5 Flash Cyber \u00abmultiple times at high speed and low cost,\u00bb allowing the AI agent to scan more code paths and find vulnerabilities.<\/p>\n

\n
\"Cybersecurity\"<\/a><\/div>\n<\/div>\n

The release of 3.5 Flash Cyber comes alongside Gemini 3.6 Flash and 3.5 Flash-Lite<\/a>, which are optimized for improved coding, knowledge work, and multimodal performance and low-latency tasks, respectively.<\/p>\n

\u00abGiven the dual-use nature of this technology, we have taken an intentional approach to how we deploy 3.5 Flash Cyber,\u00bb Raluca Ada Popa, DeepMind’s Gemini Security Lead, and Four Flynn, vice president of security and privacy at DeepMind, said<\/a> in a blog post shared with The Hacker News ahead of publication.<\/p>\n

\u00abAs part of a limited-access pilot program, 3.5 Flash Cyber will be exclusively available to governments and trusted partners via CodeMender, expanding over time. This will give frontline defenders a head start in finding and fixing critical vulnerabilities before they can be exploited, while mitigating against broader misuse.\u00bb<\/p>\n

Since 3.5 Flash Cyber runs solely inside CodeMender, it’s easy to set guardrails that enable the AI agent’s defense functions while disabling other cyber activity, the spokesperson added. This is to prevent scenarios where a model refuses to handle scenarios that prohibit defenders from performing AI-assisted forensic analysis.<\/p>\n

\"\"<\/a><\/div>\n

In evaluations conducted by the AI research laboratory, 3.5 Flash Cyber has been found to outperform Gemini 3.5 Flash and 3.6 Flash when it comes to unearthing new vulnerabilities in codebases. Additional stress-testing of the model on complex projects like Google Chrome and Apple Safari has revealed it to have \u00absignificantly\u00bb surpassed Gemini 3.5 Flash, 3.6 Flash, and Anthropic Claude Opus 4.6.<\/p>\n

\u00ab3.5 Flash Cyber consistently discovered more unique vulnerabilities compared with 3.5 Flash and Claude Opus 4.6,\u00bb it pointed out. \u00abWhen tested on the highly complex V8 JavaScript Engine across a fixed number of invocations, Gemini 3.5 Flash Cyber found 55 unique confirmed issues, compared to 47 found by Gemini 3.5 Flash and 36 found by Opus 4.6, including 10 issues that no other model caught.\u00bb<\/p>\n

\n
\"Cybersecurity\"<\/a><\/div>\n<\/div>\n

Like in the case of Anthropic and OpenAI, Google has put 3.5 Flash Cyber to the test to uncover remote code execution vulnerabilities in public APIs and a memory-corruption vulnerability in a sensitive production service. The model is also said to have produced a 100% reliable remote-code execution exploit that bypassed standard mitigation techniques like Address Space Layout Randomization (ASLR) and Write XOR Execute (W^X).<\/p>\n

Google said it’s separately bringing CodeMender’s foundational capabilities directly to customers with generally available Gemini models through the Gemini Enterprise Agent Platform<\/a>.<\/p>\n

\u00abBy powering CodeMender with 3.5 Flash Cyber, we’re providing a highly capable, scalable, and affordable architecture designed to help more defenders secure software,\u00bb it added.<\/p>\n<\/div>\n","protected":false},"excerpt":{"rendered":"

\ue804Ravie Lakshmanan\ue802Jul 21, 2026Software Security \/ Artificial Intelligence Google’s DeepMind on Tuesday announced the release of Gemini 3.5 Flash Cyber, a specialized artificial intelligence (AI) model built atop 3.5 Flash…<\/p>\n","protected":false},"author":1,"featured_media":1908,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[233,1240,1377,2653,282,2,188,588,474],"class_list":["post-1907","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-uncategorized","tag-cyber","tag-find","tag-fix","tag-flash","tag-gemini","tag-google","tag-launches","tag-software","tag-vulnerabilities"],"_links":{"self":[{"href":"https:\/\/thedigitalfortress.us\/index.php?rest_route=\/wp\/v2\/posts\/1907","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/thedigitalfortress.us\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/thedigitalfortress.us\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/thedigitalfortress.us\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/thedigitalfortress.us\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=1907"}],"version-history":[{"count":0,"href":"https:\/\/thedigitalfortress.us\/index.php?rest_route=\/wp\/v2\/posts\/1907\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/thedigitalfortress.us\/index.php?rest_route=\/wp\/v2\/media\/1908"}],"wp:attachment":[{"href":"https:\/\/thedigitalfortress.us\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=1907"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/thedigitalfortress.us\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=1907"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/thedigitalfortress.us\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=1907"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}