Gemini 3.8 Live vs Extended Thinking: Which one should you choose?

Gemini 3.8 Live is Google's latency-optimized realtime audio model. Gemini 3.8 Live Extended Thinking adds background reasoning while the conversation continues. Both use the same LiveKit AgentSession and Google RealtimeModel integration. Sign up for LiveKit Cloud: https://cloud.livekit.io/signup?utm_source=youtube&utm_medium=video&utm_campaign=devrel&utm_content=q5yF-r_CV_o Model IDs: - `gemini-3.8-live` - `gemini-3.8-live-extended-thinking` In this video, I run the same drawing agent against both models. The prompt, instructions, tool, and session code stay fixed. Then I walk through the six-run sample, show why a completed output turn can still be followed by more work, and build the LiveKit quickstart. The sample is evidence from these runs, not a general benchmark. Gemini 3.8 Live called the drawing tool in two of three sessions. Extended Thinking called it in all three and produced four to eight audible progress segments before each call. 📚 Resources 📚 - LiveKit Gemini Live API docs: https://docs.livekit.io/agents/models/realtime/plugins/gemini/?utm_source=youtube&utm_medium=video&utm_campaign=devrel&utm_content=q5yF-r_CV_o - Agent docs: https://docs.livekit.io/agents/?utm_source=youtube&utm_medium=video&utm_campaign=devrel&utm_content=q5yF-r_CV_o 🤝 Join the Community: https://community.livekit.io/?utm_source=youtube&utm_medium=video&utm_campaign=devrel&utm_content=q5yF-r_CV_o #livekit #ai #voiceai