Google rolled out Gemini 3.8 Live and Extended Thinking
Google is splitting its new live voice stack between fast conversational agents and slower multi-step reasoning.
Gemini 3.8 Live is built for low-latency speech, visual grounding, tool use and automatic switching across 97 languages. Extended Thinking is aimed at longer work, reasoning while it speaks and giving progress cues during tasks. The piece cites benchmark wins and demos ranging from camera-guided chess to React component generation from a sketched wireframe. Rollout starts through the Gemini API, AI Studio, Search Live, Gemini Live, Workspace apps and enterprise previews. TestingCatalog's note
Gemini 3.8 Live is built for low-latency speech, visual grounding, tool use and automatic switching across 97 languages. Extended Thinking is aimed at longer work, reasoning while it speaks and giving progress cues during tasks. The piece cites benchmark wins and demos ranging from camera-guided chess to React component generation from a sketched wireframe. Rollout starts through the Gemini API, AI Studio, Search Live, Gemini Live, Workspace apps and enterprise previews. TestingCatalog's note
score 8