Google introduces Gemini 3.8 Live and 3.8 Live Extended Thinking for AI agents
The models are designed to handle complex tasks with improved efficiency and visual grounding. They are part of Google's ongoing efforts to enhance AI capabilities in the agentic era.
Google has unveiled Gemini 3.8 Live and 3.8 Live Extended Thinking, two new models aimed at improving the performance of AI agents. These models are built for scale and cost efficiency, combining conversational intelligence with fluid dialogue and visual grounding. They are intended to address the growing complexity of managing AI agents, which now outnumber humans 100 to 1.
The development of these models comes as part of a broader trend in AI, where the focus is shifting toward creating more sophisticated and efficient systems. Current solutions for managing AI agents are often complex and limited in their ability to monitor API calls outside of proxy, subprocesses, or config rewrites made by agents. This limitation is becoming increasingly unsustainable as the number of AI agents continues to grow.
Gemini 3.8 Live is specifically designed to handle tasks that require both conversational intelligence and visual grounding. The model's version, 3.8, represents a significant step forward in the evolution of AI agents. This advancement is expected to have a major impact on how AI systems are developed and deployed in various industries.
The introduction of these models is likely to influence the market for AI tools and services. Companies may need to reassess their strategies for managing AI agents, considering factors such as cost, latency, and vendor lock-in. The shift toward more efficient and scalable AI systems could also lead to changes in governance and regulatory frameworks as the technology continues to evolve.
As the models continue to develop, their impact on the AI landscape will become more apparent. The ongoing refinement of Gemini 3.8 Live and 3.8 Live Extended Thinking suggests that Google is committed to advancing AI capabilities in ways that address the challenges of the agentic era. This development is part of a larger effort to create more robust and efficient AI systems that can adapt to the demands of modern applications.