Hello,
We are using Gemini 3 Flash Live for realtime voice conversations.
In the Google AI Studio usage dashboard, input tokens are being tracked correctly, but output tokens always remain empty, even after successful realtime exchanges.
We would like to understand whether this is expected behavior for Gemini 3 Flash Live, whether output tokens are tracked elsewhere, or whether this is a dashboard/reporting limitation.
At the moment, we can only see input token usage, which makes it difficult to estimate the full cost of realtime conversations accurately.
Could you please clarify this?
Thank you.