Google Launches Gemini 3.8 Live: Advances Voice AI with Near Real-Time Reasoning and Tool Execution
7
What is the Viqus Verdict?
We evaluate each news story based on its real impact versus its media hype to offer a clear and objective perspective.
AI Analysis:
The feature set represents a genuinely significant improvement in usability (the 'seamless' experience), justifying a higher impact score, but the messaging is highly competitive and incremental to existing agent concepts, keeping the hype score moderate.
Article Summary
Google announced Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking, positioning them as leading advancements in voice-based AI agents. The primary focus is mitigating the perceived latency of conversational AI by enabling near-real-time reasoning and simultaneous speech-and-thought processing. Critically, these models can execute third-party tool calls and APIs in the background without disrupting the natural flow of conversation, offering a more seamless user experience. Beyond real-time capabilities, the models boast advanced features including automatic language detection (supporting 97 languages), near real-time visual grounding, and early verbal cue recognition, making interactions feel more human-like. The models are available via APIs and Google Workspace, with clear pricing tiers outlined for both audio inputs and outputs.Key Points
- The new Gemini 3.8 Live models specialize in managing background task execution (tool calls, API integrations) while maintaining a natural, uninterrupted conversational pace.
- Technological improvements include near real-time reasoning, multi-language support (97 languages), and advanced conversational cues, enhancing the human-like feel of AI interaction.
- Google has positioned the tools for enterprise adoption, making them available through the Gemini API, Google Workspace, and emphasizing developer integration via partner platforms.

