Google launches its new AI agent, Gemini Deep Research, built for developers.
The agent runs on Gemini 3 Pro, Google’s most advanced and factual model yet.
Developers can now embed Deep Research into apps using the new Interactions API.
Deep Research autonomously gathers, analyzes and synthesizes complex data.
It plans research steps, reads sources, spots gaps, and improves accuracy.
Google says the agent reduces hallucinations during deep research tasks.
A new benchmark DeepSearchQA tests agent skills on multi-step web tasks.
Google may integrate Deep Research into Search, Gemini App and NotebookLM.
The launch lands amid a broader AI race, with OpenAI releasing GPT-5.2
This marks a key shift toward agents that reason, search and synthesize deeply.