Gemini 3.6 Cuts Costs
Google just dropped Gemini 3.6 Flash to slash your API costs! 📉 Here is how the new efficiency update changes the game for enterprise AI.
What the video says
$1.50 per million input tokens is the new baseline for high-volume enterprise AI, but can a model actually cost less while doing the exact same thinking? Google answered that on July 21st, 2026, by quietly releasing Gemini 3.6 Flash, a landmark mid-cycle update designed entirely around efficiency and slashing operational costs.
While developers were still waiting for the delayed Gemini 3.5 Pro, this disruptive release bypassed the hype to deliver a major step change in practical economics. The model achieves a 17% efficiency gain, which translates directly to a compounded cost reduction for massive workloads.
Instead of paying $9 for Gemini 3.5 Flash output, users now pay just $7.50 per million output tokens. It is not a raw intelligence leap, scoring a flat 50 on the Artificial Analysis Intelligence Index, but it still outperforms 3.5 Flash on applied benchmarks like coding and computer use.
Alongside this release, Google shipped Gemini 3.5 Flash Lite to power Search, even as they begin their most ambitious pre-training run yet for the upcoming Gemini 4.
