AiGpu

Industry·

Google Opens Gemini Tasks to Millions, Turning Chatbot into Active Agent

Google has widened availability of Gemini Tasks, a mode that lets the AI assistant make calls, book appointments, and complete purchases on behalf of users. The feature now runs on the new Gemini 4 Hydrogen model and is free for hundreds of millions of accounts.

Gemini Tasks interface showing toggle between Chat and Tasks modes on a smartphone

Google is rolling out Gemini Tasks to a vastly larger audience, shifting the assistant from a conversational tool into an agent that can act on a user's behalf. By flipping a toggle from Chat to Tasks, the system gains permission to call businesses, schedule appointments, process payments through Google Pay, and securely sign into websites when instructed.

The capability first appeared in May as an experimental "Gemini Spark" for select testers, then graduated to Google AI Ultra subscribers under the name Task Mode. Now it is shipping at no extra cost to users of the newly released Gemini 4 Hydrogen model, instantly putting agentic workflows in front of hundreds of millions of people.

Why it matters for GPU and AI infrastructure

Agentic features like Tasks demand low-latency inference at massive scale, because each user request can trigger multiple model calls, tool invocations, and real-time API interactions. This workload profile favors GPU clusters optimized for high-throughput, low-latency serving — precisely the environment AiGpu Cloud is built to deliver. As more products embed autonomous assistants, demand for elastic, high-density GPU capacity will continue to accelerate.

Google's decision to run Tasks on the Gemini 4 Hydrogen model also signals a push toward more efficient model architectures that balance capability with serving cost. For infrastructure providers, that means tracking not just raw FLOPS but also memory bandwidth, KV-cache efficiency, and scheduling overhead when sizing clusters for agentic workloads.

  • aigpu
  • ai gpu
  • ai gpu cloud
  • aigpu dubai
  • gemini
  • google
  • ai agents
  • infrastructure

By AiGpu Editorial · Editorial rewrite based on public reporting (TechCrunch AI)

← All articles