Google ships Gemini 3.6 Flash with built-in computer use and teases Gemini 4

While everyone watches the giant models, Google is playing the opposite card: small, fast, cheap models designed to be called millions of times a day inside applications and automated agents. On July 21 it launched Gemini 3.6 Flash, the update to its everyday workhorse line.

The most interesting addition is built-in computer use: the model can operate a computer directly — clicking, typing, navigating windows — with no extra components. That is the capability agents need to fill in forms, make purchases, or handle web-based paperwork on our behalf. The model is also more concise: roughly 17% fewer tokens for the same work, which means lower bills for anyone paying per token.

Alongside Flash comes 3.5 Flash-Lite, cheaper still, for simple high-volume tasks. There is a shadow over the picture, though: the flagship Gemini 3.5 Pro has been delayed by several months because internal testing showed it falling short on coding and long-horizon reasoning. Meanwhile, Google is signaling that the Gemini 4 generation is on the way.

The strategy is clear: own the tier where the real volume lives — enterprise agents running around the clock — while the race for maximum-capability models slows down and gets more expensive for everyone.