> ## Content Index > Fetch the complete content index at: https://efreedirectory.com/llms.txt > Use this file to discover other available public pages before exploring further. # Gemini 2.5 Pro and Flash Retire on 1 December, and Free Users Move Down - URL: https://efreedirectory.com/gemini-2-5-retirement-routes-to-flash/ - Published: 2026-11-04T04:22:10.000Z - Updated: 2026-11-04T04:22:10.000Z - Description: Google announced both retirements on 2 November and neither is a shutdown. The traffic simply moves, at a different price, to a model you did not choose. - Author: Chintan Dhokai - Tags: tool, AI Assistants **Gemini 2.5 Pro and Gemini 2.5 Flash were deprecated on 2 November 2026 and retire on 1 December 2026\. From that date, requests to gemini-2.5-pro route automatically to Gemini 3.8 Flash and requests to gemini-2.5-flash route to Gemini 3.5 Flash-Lite, billed at the destination model's rates. Both models have already disappeared from the model picker in Google AI Studio.**[1](#sources) Neither retirement is a shutdown. The endpoint answers, the model string is accepted, and the bill is calculated against whichever model now sits behind it. That is the quietest kind of breaking change: nothing errors, and the thing you shipped is no longer the thing you are running.[2](#sources) ## What happens to each model on 1 December | Old model | Deprecated | Routes to | Paid input / output per 1M tokens | Free tier at destination | | ------------------------- | --------------------- | --------------------- | ---------------------------------------------- | ------------------------ | | gemini-2.5-pro | 2 Nov 2026 | gemini-3.8-flash | $0.75 / $3.75 until 31 Dec, then $1.50 / $7.50 | Free | | gemini-2.5-flash | 2 Nov 2026 | gemini-3.5-flash-lite | $0.30 / $2.50 | Free | | gemini-2.5-pro (before) | was live until 30 Nov | itself | $1.25 / $10.00 up to 200k | Free | | gemini-2.5-flash (before) | was live until 30 Nov | itself | $0.30 / $2.50 | Free | **Prices measured on 27 October 2026.** Vendor plan terms, allowances and prices change without notice, so treat every figure on this page as a dated snapshot rather than a permanent fact. Where a number here differs from the vendor's own pricing page, the vendor's page is correct and the figure above is simply older than it. Read down the third column and the shape of the change becomes obvious. 2.5 Pro traffic is promoted to a newer flagship-class Flash model. 2.5 Flash traffic is demoted to Flash-Lite, which Google describes as a cost-efficient model built for high-volume agentic tasks, translation and simple data processing. If you tuned prompts against 2.5 Flash and never touched the model string, you are being moved to a cheaper, smaller sibling on a date you did not pick. No error, no notification in the response body, no deprecation header. Just quietly worse answers. ## The free tier problem nobody mentions Here is the part that matters most if you never pay Google anything. Both 2.5 models were free-tier eligible. So are both destinations, so nobody loses access outright, which is why this change has not generated the usual outrage. But free-tier access is not neutral elsewhere in the Gemini catalogue. Gemini 4 Argon, the flagship, is marked plainly as not available on the free tier. Gemini 2.5 Flash Image is not available on free either. Several of the newer generation models are paid-only or restricted. The 2.5 generation was the last wide-open door in the free tier, and both its Pro and Flash endpoints are now being closed on the same day. What replaces them on the free path is 3.8 Flash and 3.5 Flash-Lite, which are good models but sit lower in the family than the names they replace implied. ## You have twenty-seven days Google's own guidance is to update the model string before 1 December and read the migration guide. That is not a formality, because the destination models carry breaking API changes: - thinking\_level replaces thinking\_budget - Function calling and thought\_signature validation behave differently - Sampling parameters have been removed - Prefilled model turns are no longer supported Any of those will bite a production prompt chain. Prefilled turns in particular are a common trick for forcing JSON output, and code that relies on them does not degrade gracefully. Two paths forward. Change the model string yourself before 1 December and test against the destination model while you still control the timing. Or leave it and let the reroute happen, in which case the breaking changes arrive without a deploy and the first symptom is a parse error in production. The first path is the only one that lets you measure the difference between the model you chose and the model you were given. ## What this says about the Gemini free tier The Gemini free tier has been shrinking on a schedule all autumn. [Flash and Pro left the consumer free plan on 9 October](https://efreedirectory.com/gemini-free-plan-limits/). [Argon doubles in price on 1 December](https://efreedirectory.com/gemini-4-argon-price-doubles/), the same day 2.5 retires, and the 3.8 Flash promotional rate ends on 31 December. Three deadlines landing inside thirty days is not a coincidence. Google is clearing the old generation out ahead of a price reset on the new one, and the free tier is being rebuilt around Flash and Flash-Lite rather than around Pro-class access. That is a defensible strategy. It is also worth understanding before you build a project on a model string that Google has already scheduled for replacement. ## Frequently asked questions **When do Gemini 2.5 Pro and Flash stop working?** 1 December 2026\. After that date requests to both model strings route to their replacements automatically. **Will my code break immediately?** No. The model string still works, which is the problem. It just answers as a different model. **What does gemini-2.5-pro route to?** Gemini 3.8 Flash, billed at 3.8 Flash rates. **What does gemini-2.5-flash route to?** Gemini 3.5 Flash-Lite, billed at Flash-Lite rates. **Is Gemini 3.8 Flash free?** Yes, on the free tier. Its paid rate is $0.75 input and $3.75 output per million tokens until 31 December, then $1.50 and $7.50. **Do the replacements support thinking\_budget?** No. thinking\_level replaces it, alongside changes to function calling, thought signatures, sampling parameters and prefilled turns. *Deprecation dates, the 1 December reroute, destination models and breaking-change list verified in Google's Gemini API* [*deprecations page*](https://ai.google.dev/gemini-api/docs/deprecations?ref=efreedirectory.com) *and* [*release notes*](https://ai.google.dev/gemini-api/docs/changelog?ref=efreedirectory.com) *for 2 November 2026\. Free-tier availability and per-model pricing from the* [*Gemini Developer API pricing page*](https://ai.google.dev/gemini-api/docs/pricing?ref=efreedirectory.com)*, both read on 4 November 2026.* ## Sources 1. Gemini API: Deprecations: [ai.google.dev/gemini-api/docs/deprecations](https://ai.google.dev/gemini-api/docs/deprecations?ref=efreedirectory.com). Used for: Deprecation dates, the 1 December automatic reroute, destination models, the free-tier eligibility note and the breaking-change list. 2. Gemini API: Release notes, 2 November 2026: [ai.google.dev/gemini-api/docs/changelog](https://ai.google.dev/gemini-api/docs/changelog?ref=efreedirectory.com). Used for: The 2 November 2026 announcement of both 2.5 Pro and 2.5 Flash deprecations. 3. Gemini Developer API pricing: [ai.google.dev/gemini-api/docs/pricing](https://ai.google.dev/gemini-api/docs/pricing?ref=efreedirectory.com). Used for: Per-model paid rates, free-tier availability and the 3.8 Flash promotional pricing through 31 December. *All sources are the vendor's own plan or documentation pages, checked on 27 October 2026\. Plan terms change without notice, so confirm current limits before you rely on them.*