Google has launched Gemini 3.6 Flash and Gemini 3.5 Flash-Lite, expanding the Gemini AI portfolio. The new models are designed to make high-volume AI workloads, agentic tasks, and document processing faster, more affordable, and more efficient.
Tech: Google has introduced two new models in its Gemini AI platform, Gemini 3.6 Flash and Gemini 3.5 Flash-Lite. The company claims that these models will provide developers and enterprise customers with a more effective AI experience with lower token usage, faster processing, and better performance.
Two new models added to Gemini AI portfolio
Google has officially rolled out Gemini 3.6 Flash and Gemini 3.5 Flash-Lite, strengthening its AI ecosystem. According to the company, both models are specifically designed to handle high-volume AI, agentic workloads, document processing and multimodal tasks more efficiently.
The company says the new Gemini 3.6 Flash delivers better performance while using fewer output tokens than the previous Gemini 3.5 Flash. Whereas Gemini 3.5 Flash-Lite has been said to be the fastest and economical model of the entire 3.5 series.
What’s new in Gemini 3.6 Flash?
According to Google, Gemini 3.6 Flash is optimized for complex tasks like coding, data analysis, multimodal workloads, document parsing and report generation. The model uses fewer reasoning steps and fewer tool calls when completing multi-step tasks, making processing more fast and effective.
The company has claimed that in some benchmark tests this model has performed better than the earlier version. It has also added advanced security measures to prevent misuse related to sensitive areas like cyber security and CBRN (Chemical, Biological, Radiological and Nuclear).
Flash-Lite: fast speed and low cost model
Google has introduced Gemini 3.5 Flash-Lite for users and developers who need faster AI processing at a lower cost. This model is designed for agentive search, document processing, and high-throughput workloads.
In this, developers can choose different thinking levels as per their need. Low latency and cost-effective modes are available for simple tasks, while higher levels of reasoning capability are offered for complex multi-step tasks.
Cyber model also introduced
Google has also introduced a special model called Gemini 3.5 Flash Cyber. It is designed to identify, analyze, and fix security vulnerabilities in software. This model will work with Google’s CodeMender Security Agent.
However, this cyber security model will not be available to common users. The company will currently make it available only to governments and select trusted partners under a limited access pilot program.
Where will you find the new Gemini models?
According to Google, Gemini 3.6 Flash and Gemini 3.5 Flash-Lite are now available on Google AI Studio, Android Studio, Gemini API, and Gemini Enterprise platforms. Both models can also be accessed through the Gemini app, while Flash-Lite is also being rolled out in Google Search in a phased manner.
The company also revealed that Gemini 3.5 Pro is currently in the testing phase with partners and will be made widely available soon. Additionally, Google has confirmed that it is also working on the pre-training phase of the next generation Gemini 4 model, although its launch timeline has not been disclosed yet.
Comments are closed.