Agent Upstart

Cool things people are building with agents — and practical help getting the most out of them.

Playbooks · Coding tasks

Gemini 2.5 Pro and Flash retire December 1

Written November 5, 2026

Google is deprecating Gemini 2.5 Pro and Flash on December 1, with API calls rerouted. What changes for your calls, and what to check first.

What changes December 1

Both models are already deprecated and can no longer be picked in Google AI Studio. The Gemini API release notes for November 2 set the date: effective December 1, 2026, every API request to gemini-2.5-pro automatically routes to gemini-3.8-flash and is billed at Gemini 3.8 Flash rates, and every request to gemini-2.5-flash automatically routes to gemini-3.5-flash-lite and is billed at Gemini 3.5 Flash-Lite rates. The deprecations page lists December 1, 2026 as the shutdown date for both models, and notes that the dates it lists are the earliest possible dates a model might be retired.

The important part is what rerouting means in practice: calls do not fail after December 1. They are answered by a different model, at that model's price, with that model's behaviour. A request that used to reach Gemini 2.5 Pro reaches Gemini 3.8 Flash instead, and a request that used to reach Gemini 2.5 Flash reaches Gemini 3.5 Flash-Lite. Access to the 2.5 models is already limited to users who have actively used them in the past; new projects are pointed at the current models. Note that gemini-2.5-flash-lite is not deprecated and stays available through the API.

The same day carries a second change worth knowing about: Gemini 4 Argon's price doubles on December 1. See Gemini 4 Argon doubles in price on December 1.

What to change before then

  1. Find every gemini-2.5-pro and gemini-2.5-flash string. Search code, configuration files, agent settings, and any pinned model version. Change Pro calls to gemini-3.8-flash and Flash calls to gemini-3.5-flash-lite — do not wait for the rerouting to do it for you, because the rerouted call is billed and behaves as the replacement.
  2. Read the migration guides for the replacements. The Gemini 3.8 Flash migration guide and the Gemini 3.5 Flash-Lite migration guide cover the breaking API changes: function calling and thought_signature validation, thinking_level replacing thinking_budget, and removed sampling parameters and prefilled model turns. Anything that sets the old parameters needs updating, not just the model string.
  3. Recheck the cost. Rerouted calls are billed at the replacement's rates, so a budget built on 2.5 Pro or 2.5 Flash prices no longer describes what the calls cost. Compare the replacement's rates on the pricing page before December 1.
  4. Test the replacement on your own prompts. A rerouted call returns an answer from a different model. Run your usual prompts against the replacement and confirm the answers are still good enough before the switch happens on its own.

Limits

Related playbooks

Sources

Back to: Coding playbooks hub · Also: Gemini 4 Argon doubles in price on December 1 · Agent change deadlines: retirements and shutdowns