Google updated Gemini, but Pro is delayed again

Here’s a quick AI update: Google DeepMind has released three new Gemini models, but the Pro refresh is once again absent from this release.

Three points stood out to me:
— Gemini 3.6 Flash is positioned as a workhorse for coding, knowledge and multimodality, while using tokens more efficiently.
— Gemini 3.5 Flash-Lite is aimed at the lowest possible cost.
— Gemini 3.5 Flash Cyber is designed to find and fix vulnerabilities, and is available only to a limited group of partners.

Google is betting on efficiency, latency and stability for AI agents in production.

Why it matters: the market is moving fast, and practical use cases—from development to cybersecurity—are becoming the main battleground.

What matters more right now: flagship quality or efficient applied models?