Google has expanded its Gemini AI family with the introduction of Gemini 3.6 Flash and Gemini 3.5 Flash-Lite, two new models engineered to deliver faster performance, lower latency, and reduced operational costs for developers and enterprises. The company also unveiled Gemini 3.5 Flash Cyber, a specialized cybersecurity-focused model, and confirmed that development of the next-generation Gemini 4 is already underway.
Gemini 3.6 Flash, the successor to Google’s workhorse model, is designed specifically for coding, knowledge-intensive tasks, multimodality, and AI agents. It processes text, images, videos, audio, and PDFs, and offers improved coding abilities, better computer control functionality, and enhanced multimodality without latency issues. Google claims the model is more token-efficient and requires fewer reasoning steps to complete complex workflows, saving developers both time and money on inference.
For high-volume AI workloads, Google launched Gemini 3.5 Flash-Lite, a cost-effective version of the Flash series ideal for tasks such as document parsing, classification, data extraction, translation, and large-scale automation. Flash-Lite supports multimodal inputs and provides fast responses, making it suitable for building lightweight AI agents or processing high-volume data.
Alongside these general models, Google announced Gemini 3.5 Flash Cyber, a dedicated cybersecurity model designed to detect software vulnerabilities, ensure code security, and help remediate threats. Initially available only to governments and trusted partners, the model aims to speed up security assessments and reduce costs associated with repetitive code scans.
Both Gemini 3.6 Flash and Gemini 3.5 Flash-Lite are now available through Google AI Studio, Gemini API, Android Studio, and other Google enterprise AI solutions. Google has also begun deploying the Flash-Lite version in consumer apps such as Google Search.
While expanding its lightweight AI portfolio, Google has not yet provided a release date for the long-awaited Gemini 3.5 Pro model. Originally expected earlier this year, the model is still undergoing performance testing for coding tasks. According to Google, its release will follow additional validation work, while development of Gemini 4 has already started.