Hello,
Thanks to Gemini, I’ve been able to create a completely new kind of game that wasn’t possible before.
However, due to the nature of the game, API latency is extremely important — and recently it seems noticeably slower than before.
I am currently sending audio input to the Gemini 2.5 Flash (without thinking) model and receiving text output.
Our measured average API latency is:
-
Aug 5: 1.21 sec
-
Dec 9: 2.29 sec
Has there been any recent change that might have caused the API to become slower?
Additionally, are there any plans to improve performance or restore the previous latency levels in the future?
Thank you.