AI library

Source · release

Gemini API changelog: Gemini 2.5 Pro and 2.5 Flash deprecated

Author
Google
Published
2 November 2026
Added
3 November 2026
Original
https://aistudio.google.com/docs/changelog

Summary

On 2026-11-02 Google deprecated gemini-2.5-pro and gemini-2.5-flash. They can no longer be selected in Google AI Studio, and from 2026-12-01 API requests to them are routed automatically to newer models and billed at the newer models' rates. The Gemini API docs now live on aistudio.google.com (ai.google.dev redirects there).

Key points

  • gemini-2.5-pro routes to gemini-3.8-flash from 2026-12-01, billed at 3.8 Flash rates. A Flash model replaces a Pro model.
  • gemini-2.5-flash routes to gemini-3.5-flash-lite from 2026-12-01, billed at 3.5 Flash-Lite rates.
  • The deprecations page lists 2026-12-01 as both the routing and shutdown date and calls the dates the earliest possible. API access was already limited to existing users.
  • Breaking changes to plan for: function calling and thought_signature validation, thinking_level replacing thinking_budget, removed sampling parameters, and no prefilled model turns.
  • Checked on the same day: the pricing page (2026-11-03) bills Computer Use "as regular tokens per model pricing" and does not offer it on the free tier. The page no longer lists gemini-3.5-flash, and the deprecations page still does not say how its traffic routed to 3.6 Flash is billed.

Why it matters

Code still pinned to 2.5 will silently switch model and price on 2026-12-01, and may break on the API changes; migrate explicitly before then.