Google deprecates Gemini 2.5 Flash, routing its API traffic to Gemini 3.5 Flash-Lite
Google says gemini-2.5-flash is deprecated and will automatically route to gemini-3.5-flash-lite, billed at Flash-Lite rates, starting December 1, 2026.
Google has deprecated gemini-2.5-flash. From December 1, 2026, API requests that still name that model will be sent to gemini-3.5-flash-lite and billed at Gemini 3.5 Flash-Lite rates. Google's Gemini API changelog for November 2, 2026 says the model is no longer available to select in Google AI Studio. Requests that keep the old model string will be served by a different model, billed at Flash-Lite rates.
The changelog gives developers a window to avoid that. Before December 1, 2026, they can change their model string to gemini-3.5-flash-lite and review Google's Gemini 3.5 Flash-Lite migration guide. Requests that are not updated will be routed automatically to the new model. The changelog lists the breaking changes that come with that move: function calling and thought_signature validation, thinking_level replacing thinking_budget, and removed sampling parameters and prefilled model turns. It presents them as a single list rather than explaining each one.
The changelog does not say how Gemini 3.5 Flash-Lite compares with 2.5 Flash in output quality or in price, beyond stating that requests will be billed at its rates. A developer reading only this document cannot tell whether the new model will cost more or less per request. The migration guide the changelog links to is the place to look, and the changelog does not summarize it.
What changed
gemini-2.5-flash is deprecated, and its API requests are routed to gemini-3.5-flash-lite and billed at Gemini 3.5 Flash-Lite rates.
Type: Breaking change · Takes effect: Tuesday, December 1, 2026 · Applies to: gemini-2.5-flash, gemini-3.5-flash-lite
Why it matters
Applications built on gemini-2.5-flash will move to a different model and price tier after December 1, so output behavior and cost can change without a code update if they do not migrate.
What you can do
Change the model string to gemini-3.5-flash-lite before December 1, 2026, and check the Flash-Lite migration guide for the breaking changes.
How the story developed
The official changelog adds the list of breaking API changes, including function calling, thought_signature validation, thinking_level replacing thinking_budget, and removed sampling parameters and prefilled model turns.
Original source
Where we saw it
Gemini API changelog
AI wrote this summary from the sources listed, and nobody reviewed it individually before publication. Check the original source before you rely on it. How it works