Google Launches Gemini 3.8 Live with Deep Reasoning Capabilities and Real-Time Voice Interaction Across 97 Languages

Google announced the launch of its artificial intelligence models Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking, in a step representing a qualitative leap in multimodal language models, as both models combine instant voice interaction and advanced multi-step reasoning capabilities within a single integrated platform.

The standard Gemini 3.8 Live version is distinguished by its ability to conduct natural real-time voice conversations with fast, low-latency responses, while the Extended Thinking version adds a layer of deep reasoning that enables the model to process complex problems requiring intricate deductive sequences prior to answering, thereby enhancing accuracy in specialized and analytical tasks.

Both models fully support 97 languages, placing Google in an advanced position in the multilingual AI applications market, particularly in emerging markets that have historically lacked adequate representation in major commercial AI models.

These releases come amidst intensifying competition among Google, Anthropic, and OpenAI for leadership in the large AI model market, with each entity betting on combining speed and analytical depth as the primary determinant for capturing market share over the medium term.

What do these terms mean?

Real-time Voice Interaction: The model's ability to understand speech and respond to it instantly without noticeable latency, enabling continuous natural conversations.

Multi-step Reasoning (Extended Thinking): An approach in which the model breaks down a complex problem into sequential steps, analyzing them one by one before reaching a final answer, thereby improving reasoning quality in difficult tasks.