Steps to reproduce:
- Begin speaking and ask the model a question using voice input.
- While the model is responding, attempt to interrupt it with additional input or to correct it.
Expected result:
The model should stop speaking immediately when interrupted, allowing new input to be processed with minimal delay.
Actual result:
The model continues to speak for up to 30 seconds before adjusting to the interruption, causing a delay in processing the new input.
Additional information:
Browser: Chrome
OS: Windows 11