AI Pulse · Business Models & Winners

Google has launched Gemini 3.8 Flash at the same API price as its predecessor, $0.75 per million input tokens and $3.75 per million output tokens. That rate is only introductory pricing through December 31. From January 1, 2027, the standard price moves to $1.50 for input and $7.50 for output tokens per million.

For a business using the model internally, this is manageable if you know it is coming. For an agency, software company, or consultancy quoting fixed-price AI work, it matters much more. A workflow that costs $500 a month to run today could cost roughly twice that after the introductory period, assuming usage stays the same. That can erase margin quickly if your client contract locks in a price for 12 months.

The practical move is simple. Separate model usage costs from your service fee where possible, or build a pricing review clause into longer client agreements. Also check whether your AI tools are using Gemini behind the scenes, since the API cost may reach you through a vendor rather than a direct Google bill.

This is the kind of thing we build into an AI command centre, tracking the operational costs behind AI workflows so pricing changes do not become a surprise in next quarter’s numbers.