Gemini 3.8 Flash is live. Google kept token prices unchanged, but per-task costs may rise

Google introduced Gemini 3.8 Flash and the specialized Gemini 3.8 Flash Cyber on September 2, 2026. It is the company’s third Flash release in six weeks. The more important change is how the model works: Gemini 3.8 Flash performs more reasoning steps on difficult tasks and can call tools iteratively.

The per-token price is unchanged

During the introductory period, Gemini 3.8 Flash costs $0.75 per million input tokens and $3.75 per million output tokens, the same introductory rates as Gemini 3.7 Flash. Google says this pricing expires on December 31, 2026. From January 1, 2027, the listed rates are $1.50 per million input tokens and $7.50 per million output tokens.

Why can a task still cost more?

A token rate and the total cost of a task are not the same thing. If the new model uses more reasoning steps and produces more output tokens on a difficult job, the final bill for that job can rise even when the unit price is unchanged. For businesses, cost per workflow is therefore a more useful metric than the headline price per million tokens.

Where is Gemini 3.8 Flash available?

Developers can access the model through the Gemini API and Google AI Studio, enterprises through Gemini Enterprise, and Google AI Pro and Ultra subscribers through selected Google products. Gemini 3.8 Flash Cyber is restricted to trusted defenders through Google’s Fairwind Program.

The practical takeaway

Gemini 3.8 Flash shows how quickly the model market is moving. For companies using model APIs, a useful comparison should include the cost of one real task, latency and output quality. Those three measures reveal more than the published token rate alone.

Sources

Google: Introducing Gemini 3.8 Flash and 3.8 Flash Cyber The Verge: Google says its new Gemini 3.8 Flash model ‘works harder’ but might cost more