Lates News

Google launched three new Gemini AI "Flash" models on Tuesday, all focused on improving efficiency. This release comes at a time when the use of open-source models that can run at lower costs is increasing dramatically. The focus of AI competition has shifted from who can achieve the highest benchmark test scores to who can provide the best performing models at the lowest cost. The new models released include: Gemini 3.6 Flash: an upgraded version of 3.5 Flash that is faster and lower cost. Google stated that this model reduces up to 17% of output tokens while improving code writing, inference capabilities, and multimodal performance. Gemini 3.5 Flash-Lite: a lower cost model designed for high-frequency AI agents and document processing tasks. Gemini 3.5 Flash Cyber: a model focused on network security for identifying and fixing software vulnerabilities. Currently, this model is only available to government agencies and some partners through Google's CodeMender platform.