Google DeepMind has launched three new Gemini models designed for AI agents: Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber. These models prioritize efficiency, low latency, and reliability for scalable agentic workflows.

Gemini 3.6 Flash is positioned as a versatile model, offering improved coding, knowledge work, and multimodal performance. It boasts a 17% reduction in output token usage compared to 3.5 Flash, at a lower cost per token, making AI agents more cost-effective. It also shows performance gains in areas like coding precision and ML research.

Gemini 3.5 Flash-Lite is the fastest and most cost-effective in the 3.5 series, delivering 350 output tokens per second. It's optimized for high-throughput tasks such as agentic search and document processing, offering a strong price-to-performance ratio. Gemini 3.5 Flash Cyber, fine-tuned for cybersecurity, is integrated with the Code Mender agent to efficiently find and fix code vulnerabilities, initially available to governments and trusted partners.