Setting change

Gemini API: 2.5 Pro and 2.5 Flash change model on 1 Dec

From 1 December 2026, Gemini API calls to gemini-2.5-pro go to gemini-3.8-flash and calls to gemini-2.5-flash go to gemini-3.5-flash-lite.

Date
Works with
Any agent

What’s new

Google deprecated two Gemini API models on 2 November 2026. Google's changelog says: "Effective December 1, 2026, all API requests to gemini-2.5-pro will automatically route to gemini-3.8-flash and be billed at Gemini 3.8 Flash rates". For gemini-2.5-flash it says that requests "will automatically route to gemini-3.5-flash-lite and be billed at Gemini 3.5 Flash-Lite rates".

You can no longer pick either model in Google AI Studio. The deprecations page lists 1 December 2026 as the shutdown date of both. It also says access to the 2.5 models is limited to users who used them in the past. gemini-2.5-flash-lite "is not deprecated and remains available through the API until further notice".

Why it matters

This change is in the Gemini API. It affects scripts, apps and MCP servers that call the Gemini API with these model names. It is not a Gemini CLI release. If you typed one of these model names into a setting yourself, check it.

After 1 December your code still runs, but a different model answers, at a different price. Google lists breaking API changes for the new models: function calling and thought_signature validation, thinking_level in place of thinking_budget, removed sampling parameters and no prefilled model turns. Code that uses these can fail or behave differently.

How to try it

Search your own code and settings for gemini-2.5-pro and gemini-2.5-flash. Google says to change gemini-2.5-pro to gemini-3.8-flash and gemini-2.5-flash to gemini-3.5-flash-lite before 1 December. Read the migration guides that the deprecations page links, then test your code with the new name. Check the prices of the new models on Google's pricing page.

Source

Checked against the source on 3 Nov 2026

All updates