Google announced the release of three new AI models on Tuesday, all built on the Gemini 3.5 Flash foundation. The updates focus on improved efficiency and specialized capabilities.
The new lineup includes Gemini 3.6 Flash, described as a workhorse model with gains in coding and multimodal tasks. It uses up to 17 percent fewer tokens than its predecessor at a lower cost.
Gemini 3.5 Flash-Lite delivers 350 output tokens per second and supports computer use for agentic workflows. Gemini 3.5 Flash Cyber targets cybersecurity, working with an infrastructure agent called CodeMender to identify and patch vulnerabilities.
The cybersecurity model will initially be limited to governments and trusted partners. Gemini 3.5 Pro remains in testing with partners and will launch when ready, while pretraining for Gemini 4 has already begun.
