Google officially introduced the Gemini 3.6 Flash and 3.5 Flash-Lite models before the highly anticipated 3.5 Pro version. The company also shared valuable information about future model studies.

The newly announced Gemini 3.6 Flash offers higher token efficiency in line with developer and user feedback. This model completes multi-step workflows faster by consuming 17% less output tokens compared to the 3.5 Flash version.

Gemini 3.6 Flash and Flash-Lite Features

Gemini 3.6 Flash promises fewer artifacts and higher precision in coding performance. The information cut-off date of the model has been updated to March 2026.

Designed for missions that require high process volume and low latency, Gemini 3.5 Flash-Lite stands out especially in search and document processing processes. This model offers much higher performance and quality than the previous 3.1 Flash-Lite version.

Google also introduced the Gemini 3.5 Flash Cyber ​​​​model to detect and fix security vulnerabilities. This model was developed to address code security challenges in a scalable manner.

Access and Future Plans

Gemini 3.6 Flash and 3.5 Flash-Lite models are available via the Gemini application as of today. Developers can access these models through Google Antigravity, AI Studio and Android Studio.

The Gemini 3.5 Flash Cyber ​​​​model is being used with a limited pilot program in its initial phase only for government institutions and reputable partners. The company takes advantage of this model’s security spies in its CodeMender tool.

Google stated that the Gemini 3.5 Pro model is in the testing phase with partners and will be widely released when it is ready. The DeepMind team announced that they have already started working on Gemini 4, the fastest pre-training process.

How do you think these newly announced models will affect competition in the world of artificial intelligence?