Google's AI Models Struggle to Match Industry Leaders

Google recently launched two new voice AI models called Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking, touting impressive features such as real-time visual processing, 97-language support with seamless switching, and a thinking mode that reasons out loud while working.

The models are designed for natural conversation and are intended to compete with other leading voice AI models like OpenAI's GPT-4 and Anthropic. However, despite the impressive specs, Google still trails behind its competitors in terms of industry leadership.

One area where Google excels is multimodal input processing, allowing users to analyze video clips or transcribe podcasts in real-time. This feature is unmatched by other major models like Claude and GPT-5.

Despite the technical advancements, Google's AI division still faces a positioning problem - its products are strong everywhere but dominant nowhere.