Google launches three new Gemini models, focused on efficiency and agents
Gemini 3.6 Flash and two more variants, designed to lower the cost and latency of high-volume automations. Cheaper models = cheaper tasks.
Google released three new Gemini models —3.6 Flash, 3.5 Flash-Lite and 3.5 Flash Cyber— geared toward efficiency, low latency and reliability for those building AI agents at scale. Notably, it did not include the expected Gemini 3.5 Pro, apparently delayed over performance issues.
3.6 Flash is the “workhorse” model, with better code and up to 17% less token usage than its predecessor, which makes it cheaper. Flash-Lite is the most affordable in its class, for high-volume automation, and Flash Cyber specializes in finding and fixing vulnerabilities (limited access, in pilot).
Why it matters to you
Cheaper, faster models lower the cost per task. That lets you run high-volume automations —sorting messages, answering inquiries, following up— without costs spiraling as you grow.