Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber

Google has introduced three new Gemini models to improve AI agent development: Gemini 3.6 Flash, which reduces output token usage by up to 17% compared to its predecessor while offering better coding and multimodal performance at lower costs; 3.5 Flash-Lite, the fastest and most cost-effective option delivering 350 output tokens per second; and 3.5 Flash Cyber, a specialized cybersecurity model paired with CodeMender for code security applications. These models are designed to provide developers with higher efficiency, lower latency, and more reliable performance for scaling production AI agents, with Gemini 3.5 Pro also in testing and Gemini 4 currently in pre-training.