Google's Gemini 3.8 Brings Real-Time Transparency to AI Decision-Making

Google's Gemini 3.8 model has been updated to provide real-time transparency into its decision-making process, allowing users to hear the model narrate its thought process as it unfolds.

This new capability is part of Google's Extended Thinking feature, which enables the model to reason and speak simultaneously while handling complex workflows without breaking conversational flow.

The model uses early verbal cues, such as 'Let me check that,' to acknowledge user prompts while executing background tasks in real-time. This allows users to gain a better understanding of the model's limitations and path to a conclusion.

Extended Thinking has been benchmarked against other AI systems, with impressive results: 82.6 on the Artificial Analysis Speech-to-Speech Quality Index, 97.7% on Big Bench Audio, and 68.6% on τ-Voice agentic task completion.