Gemini 2.5 Pro and Flash API calls reroute to newer models on 1 December

Gemini 2.5 Pro and Flash API calls reroute to newer models on 1 December

2 November 2026: Google has deprecated two of the most widely used Gemini models on its developer API. Its release notes for 2 November say that gemini-2.5-pro and gemini-2.5-flash can no longer be picked in Google AI Studio, and that from 1 December 2026 every API request naming either model will be routed to a newer one and billed at that model's price. Requests to gemini-2.5-pro will go to gemini-3.8-flash; requests to gemini-2.5-flash will go to gemini-3.5-flash-lite. The deprecation page adds that API access to the 2.5 models is already limited to projects that have used them before. gemini-2.5-flash-lite is not deprecated and stays available until further notice.

Not yet independently verified. This rests on Google's own Gemini API changelog, deprecation, pricing and logging pages; no independent report was found. The routing applies to the Gemini API; Vertex AI has its own retirement dates, which we covered separately. We will update this when it can be confirmed, and remove this note.

Requests that will stop behaving the same way

The reroute is not a drop-in swap. Google's notes list breaking changes for both target models: function calling and thought_signature validation work differently, thinking_level replaces thinking_budget, sampling parameters are removed, and prefilled model turns are no longer accepted. Code that sets a thinking budget, tunes temperature or top-p, or starts the model's reply with a prefilled turn should be tested against the new model before 1 December rather than left to be rerouted. Google points to a migration guide for each target model and asks developers to change the model string themselves before the date.

What the bill does after the switch

Google's pricing page shows the money moving in different directions for the two models. Gemini 2.5 Pro costs $1.25 per million input tokens and $10.00 per million output tokens for prompts of up to 200,000 tokens on the paid tier. Gemini 3.8 Flash costs $0.75 input and $3.75 output until 31 December 2026, then $1.50 input and $7.50 output from 1 January 2027. So rerouted Pro traffic gets cheaper in December, and from January its input price rises above what 2.5 Pro charged for prompts of that size while output stays below it.

For Flash users the text price barely moves: Gemini 2.5 Flash costs $0.30 input and $2.50 output, and Gemini 3.5 Flash-Lite also costs $0.30 for text, image, video or audio input and $2.50 for output. Audio input, which 2.5 Flash prices at $1.00, becomes cheaper.

Requests are now stored by default on new projects

A separate change from 26 October affects data handling. Since that date, new paid-tier projects that use the Generate Content API (generateContent, streamGenerateContent and batchGenerateContent) store requests by default, as the Interactions API already did. Projects created before 26 October keep their existing setting, which is off unless logging was turned on in AI Studio. Google's logging page says stored logs expire after a default of 55 days, which a project can cut to 7, 14 or 28 days, while logs saved into a dataset do not expire. Teams with data-handling commitments can set store=false per request, which needs recent versions of Google's SDKs, or switch logging off for the project in the AI Studio settings panel.

What to check before December

• Search code and configuration for the strings gemini-2.5-pro and gemini-2.5-flash, including in batch jobs and third-party tools that call the API for you.
• Remove thinking_budget, sampling parameters and prefilled turns, or move to thinking_level, and rerun evaluations on the target model.
• Recalculate the January 2027 bill for Pro traffic at the 3.8 Flash rate.
• For projects created on or after 26 October, decide whether requests should be stored, and set the retention window.

Sources

• Google AI for Developers: Gemini API release notes, 2 November 2026 entry, 2 November 2026
• Google AI for Developers: Gemini API deprecations, 2 November 2026
• Google AI for Developers: Gemini Developer API pricing, 4 November 2026
• Google AI for Developers: Logs and datasets, 26 October 2026