Google will move Gemini 2.5 Pro calls to a different model on 1st Dec, then double that model's price a month later
On 2nd Nov'26 Google said that from 1st Dec'26 requests to gemini-2.5-pro will be routed to Gemini 3.8 Flash and billed at its rates, and requests to gemini-2.5-flash will be routed to Gemini 3.5 Flash-Lite. Google's pricing page shows Gemini 3.8 Flash at $0.75 per million input tokens and $3.75 per million output tokens until 31st Dec'26, rising to $1.50 and $7.50 from 1st Jan'27, with the same doubling on every tier. Gemini 4 Argon doubles sooner, from $2 to $4 per million input tokens and from $10 to $20 per million output tokens on 1st Dec'26.
What it actually means
Most deprecations break something, which at least makes them visible, but this one does not, because a call to gemini-2.5-pro will still return an answer on 2nd Dec'26, only from a different model with its own habits. Teams that pinned 2.5 Pro to keep outputs stable get a model swap they did not test, and a month later the bill for the model they were moved to doubles.
Before 1st Dec, find every place that calls the two retiring names, including model settings inside frameworks and old scripts, and run your own evaluation set against 3.8 Flash. Then price your volume at the January rates rather than today's, because at double the cost 3.8 Flash may no longer be the cheap option you compare it with.
From 1st Dec your Gemini 2.5 Pro calls answer from a model you have not tested, and that model doubles in price a month later.
2xGemini 3.8 Flash's price from 1st Jan'27
Source: Google AI for Developers · Friday, 6 November 2026