Gemini 3.8 Live models now available on AI Gateway
By Steven Van ·
Gemini 3.8 Live adds real-time audio and 97-language switching, while the Extended Thinking variant narrates multi-step reasoning during a live call.
Google's Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking models are now available on Vercel's AI Gateway, adding real-time spoken interaction to the platform's model lineup for voice assistants and other audio-responding apps.
- google/gemini-3.8-live supports real-time audio, visual grounding, automatic switching across 97 languages, and background tool calls that run while the conversation continues.
- google/gemini-3.8-live-extended-thinking adds multi-step reasoning that runs in parallel with speech, so it can acknowledge a request and narrate progress without interrupting the conversation.
Both models work through the AI SDK's realtime API: install the Gateway provider and a WebSocket client, mint a short-lived token, then use the model adapter to serialize and parse realtime events. Both can also be tried directly in the model playground. See the realtime quickstart for the full setup.
Vercel
The platform for frontend developers — deploy, preview, and scale web apps and AI agents with zero config.
View Vercel →