Gemini 2.5 Flash audio response latency doubled / tripled after 3.0 Pro release?

[icapora]

Hey @Richard_Wong,
yes, we are experiencing the same behavior. When sending all the audio at end_of_speech, the response shows a noticeable delay.

In this post I shared a small workshop where I explain the issues, pros, and cons of using audio streaming vs. end_of_speech, in case it helps anyone as an example or reference while debugging this behavior:

Gemini Live API: Delays or Missing input_audio_transcription Events

I created a small public repo to reproduce and document a behavior I identified while testing the Gemini Live API. When sending audio in a continuous streaming session for long periods of time, the input_audio_transcription tends to degrade or eventually stops arriving. However, when sending the full audio chunk at once (non-streaming) at the end of the user’s speech, the transcriptions remain accurate. This is consistent with what I’m also seeing in AI Studio, so I assumed this might be expect…

Hope it’s useful!