Google says its new Gemini 3.8 Flash model ‘works harder’ but might cost more
The Verge Stevie Bonifield ● Covered by 3 sources
Google just launched Gemini 3.8 Flash, a few weeks after 3.7 Flash. It keeps the same starting price, but it may burn more tokens to do the job.
Based on reporting by The Verge, Stevie Bonifield — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Google has moved fast again: Gemini 3.8 Flash is out only a few weeks after Gemini 3.7 Flash. The pitch is simple enough. This version “works harder” on tougher tasks by taking more reasoning steps and calling tools again and again until it gets where it wants to go.
On paper, the pricing hasn’t changed. Google says 3.8 Flash starts at $0.75 per million input tokens and $3.75 per million output tokens, the same as 3.7 Flash. But the company is also warning that the model may use more tokens to squeeze out better results, especially when users turn up the effort level.
That makes this less like a price cut and more like a tradeoff dressed up as an upgrade. If the model spends longer thinking, the bill can climb even when the sticker price stays put. Google is at least being blunt about that, which is refreshing in a field that usually prefers glossy promises.
Developers who care more about token use than performance can stick with Gemini 3.7 Flash. That option matters, because the new model’s gains may come with a bit of hidden appetite. And in AI, hidden appetite is often where the real cost lives.
My take — AI-written commentary, not fact-checked reporting
This is the kind of product move that should make developers squint. Same starting price, but maybe more tokens, is not exactly a gift from the efficiency gods. Google is basically telling everyone that “better” may also mean “more expensive,” which is the most honest sentence in the whole pitch.
Read more about this at: The Verge
