Google has set a firm end date for two of the most widely used models in its developer platform. In the Gemini API release notes dated November 2, 2026, the company says gemini-2.5-pro and gemini-2.5-flash are deprecated and can no longer be selected in Google AI Studio. From December 1, 2026, every API request that still names either model will be answered by a different, newer model and charged at that model's prices.
The two models do not share a successor. Calls to gemini-2.5-pro will route to gemini-3.8-flash, which Google's documentation calls its most intelligent Flash model, aimed at long-horizon software engineering, autonomous agents and complex enterprise workflows. Calls to gemini-2.5-flash will go to gemini-3.5-flash-lite, a cheaper model that Google pitches at high-volume agentic tasks, translation and simple data processing. The deprecations page now lists December 1, 2026 as the shutdown date for both 2.5 models and adds that API access to them is limited to existing users who have actively used them.
The redirect is not a drop-in swap. Google's changelog warns of breaking API changes for both paths: stricter validation around function calling and thought signatures, the thinking_level setting replacing the older thinking_budget, and the removal of sampling parameters and prefilled model turns. The Gemini 3.8 Flash migration guide spells out what that means in practice. Developers must drop temperature, top_p, top_k and candidate_count, choose a thinking level of low, medium or high (the minimal level is not supported), handle multi-turn state differently, and make sure function responses carry the required identifiers. Code written for 2.5 that relies on any of those features may fail or behave differently once the routing takes effect.
Billing shifts too. On Google's pricing page, paid-tier Gemini 2.5 Pro costs $1.25 per million input tokens and $10 per million output tokens for prompts up to 200,000 tokens. Gemini 3.8 Flash is on an introductory rate of $0.75 for input and $3.75 for output through December 31, 2026, rising to $1.50 and $7.50 on January 1, 2027. Rerouted 2.5 Pro traffic would therefore be cheaper per output token at both rates, but somewhat dearer per input token once standard pricing begins. For 2.5 Flash users the listed text rates stay level, at $0.30 for input and $2.50 for output on 3.5 Flash-Lite, while audio input drops from $1.00 to $0.30.
Not every 2.5 model is affected. The deprecations page says gemini-2.5-flash-lite is not deprecated and remains available until further notice. Several older preview versions of 2.5 Pro and 2.5 Flash had already been shut down earlier, and a 2.5 computer-use preview reached its shutdown date in July.
Automatic rerouting is a notable choice. Rather than letting calls to a retired model fail with an error, Google will keep applications running on a model the developer never selected. That avoids sudden outages for teams that miss the notice, but it also means output quality, latency, token usage and costs can change without any code change on the customer's side. Teams that pin model versions for evaluation or compliance reasons will want to test their prompts and tool calls against the replacements before the switch.
Google's notice leaves some questions open. It does not say how much traffic still runs on the two models, why 2.5 Pro maps to a Flash model rather than a higher tier, or whether customers can opt out of the automatic routing. The capability descriptions of Gemini 3.8 Flash and 3.5 Flash-Lite are Google's own, and how the replacements compare with 2.5 models on any given workload will depend on testing by each developer.