Gemini 2.5 Pro and 2.5 Flash are deprecated: requests reroute on Dec 1

Google’s Gemini API release notes, dated November 2, 2026, mark Gemini 2.5 Pro and Gemini 2.5 Flash as deprecated. If you still call them, your code will keep working after December 1, but on a different model.

What to know

• Gemini 2.5 Pro: from December 1, 2026, requests to gemini-2.5-pro automatically route to gemini-3.8-flash, billed at that model’s rates.
• Gemini 2.5 Flash: from the same date, requests to gemini-2.5-flash route to gemini-3.5-flash-lite, billed at its rates.
• Breaking changes: Google lists changes to function calling and thought_signature validation, says thinking_level replaces thinking_budget, and notes that some sampling parameters and prefilled model turns are removed. Test before the switch rather than after.
• What Google recommends: update the model string in your code before December 1.

What to do now

1. Search your code and configs for gemini-2.5-pro and gemini-2.5-flash.
2. Pick a replacement on purpose. Google’s new Gemini 4 Hydrogen (GA, $1.50 / $7.50 per million tokens) is aimed at everyday coding and agent work; compare it with Carbon and Argon in our Gemini 4 model guide.
3. Rerun your own test prompts and compare outputs and cost. Our prompt workflow shows a simple way to do it.

Check Google’s release notes for the full list of changes, and see the model tracker for current Gemini models.

Source: Google AI Studio release notes. We summarise in our own words; check the source for the full details.