{"id":5678,"date":"2026-07-23T13:47:25","date_gmt":"2026-07-23T08:02:25","guid":{"rendered":"https:\/\/blog.eastlink.com.np\/?p=5678"},"modified":"2026-07-23T14:02:03","modified_gmt":"2026-07-23T08:17:03","slug":"google-launches-gemini-3-6-flash-and-3-5-flash-lite-to-reduce-enterprise-ai-agent-costs","status":"publish","type":"post","link":"https:\/\/blog.eastlink.com.np\/?p=5678","title":{"rendered":"Google Launches Gemini 3.6 Flash and 3.5 Flash-Lite to Reduce Enterprise AI Agent Costs"},"content":{"rendered":"
<\/div>

Google has introduced Gemini 3.6 Flash<\/strong> and Gemini 3.5 Flash-Lite<\/strong>, two new AI models aimed at improving efficiency for enterprise AI agents. The company says Gemini 3.6 Flash generates 17% fewer output tokens<\/strong> than the previous version, helping reduce operational costs and latency for organizations running large-scale automated workflows.<\/p>\n

According to Google, Gemini 3.6 Flash achieved a 49% success rate on the DeepSWE benchmark<\/strong>, compared with 37% for Gemini 3.5 Flash. On the MLE Bench benchmark, performance increased from 49.7% to 63.9%<\/strong>, while its GDPval-AA v2 score improved from 1349 to 1421<\/strong>.<\/p>\n

Several companies, including Figma, Harvey, and Hebbia<\/strong>, have integrated the model into their workflows. These organizations use it for tasks such as software prototyping, document analysis, financial filing reviews, and report generation.<\/p>\n

Google also announced a built-in computer-use tool<\/strong> for Gemini models, allowing AI systems to interact directly with operating systems without requiring custom intermediary software. The company reported an OSWorld-Verified score of 83.0%<\/strong>, up from 78.4%.<\/p>\n

For high-volume tasks, Gemini 3.5 Flash-Lite<\/strong> offers lower-cost processing with pricing of $0.30 per million input tokens<\/strong> and $2.50 per million output tokens<\/strong>. Google states that the model can generate approximately 350 output tokens per second<\/strong>, making it suitable for document processing and agentic search workloads.<\/p>\n

\"\"<\/p>\n

In cybersecurity, Google introduced Gemini 3.5 Flash Cyber<\/strong>, a restricted model designed for vulnerability validation and remediation. Access is currently limited to governments and vetted partners through a pilot program. The model is used within Google\u2019s CodeMender<\/strong> security agent, where multiple AI instances review vulnerabilities and produce remediation recommendations for human approval.<\/p>\n

Google also confirmed that Gemini 3.5 Pro<\/strong> remains in partner testing and that development of the future Gemini 4 architecture<\/strong> is already underway.<\/p>\n

\n","protected":false},"excerpt":{"rendered":"

Google has introduced Gemini 3.6 Flash and Gemini 3.5 Flash-Lite, two new AI models aimed at improving efficiency for enterprise AI agents. The company says Gemini 3.6 Flash generates 17% fewer output tokens than the previous version, helping reduce operational costs and latency for organizations running large-scale automated workflows. According to Google, Gemini 3.6 Flash […]<\/p>\n","protected":false},"author":3,"featured_media":5679,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1035,1038,1036,1070,1065,1082],"tags":[],"class_list":["post-5678","post","type-post","status-publish","format-standard","has-post-thumbnail","category-ai-basics","category-ai-tools-applications","category-ethics-trends","category-learning-resources","category-tech-insights","category-technology"],"_links":{"self":[{"href":"https:\/\/blog.eastlink.com.np\/index.php?rest_route=\/wp\/v2\/posts\/5678","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/blog.eastlink.com.np\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/blog.eastlink.com.np\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/blog.eastlink.com.np\/index.php?rest_route=\/wp\/v2\/users\/3"}],"replies":[{"embeddable":true,"href":"https:\/\/blog.eastlink.com.np\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=5678"}],"version-history":[{"count":3,"href":"https:\/\/blog.eastlink.com.np\/index.php?rest_route=\/wp\/v2\/posts\/5678\/revisions"}],"predecessor-version":[{"id":5683,"href":"https:\/\/blog.eastlink.com.np\/index.php?rest_route=\/wp\/v2\/posts\/5678\/revisions\/5683"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/blog.eastlink.com.np\/index.php?rest_route=\/wp\/v2\/media\/5679"}],"wp:attachment":[{"href":"https:\/\/blog.eastlink.com.np\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=5678"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/blog.eastlink.com.np\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=5678"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/blog.eastlink.com.np\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=5678"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}