WEB DESK: Google on Thursday unveiled an upgraded version of its Gemini Deep Research agent, now powered by the company’s state-of-the-art Gemini 3 Pro foundation model.
The new Gemini Deep Research enables developers to integrate Google’s advanced research capabilities into their own applications via the newly introduced Interactions API, offering greater control in the emerging era of agentic AI.
Designed to synthesise large volumes of information and handle extensive context, the tool is already being used by customers for tasks ranging from due diligence to researching drug toxicity.
Google announced plans to soon bring the deep research agent to Google Search, Google Finance, the Gemini App, and NotebookLM, signaling a shift toward AI agents taking over information retrieval and search functions traditionally performed by humans.
The company stated that Deep Research leverages Gemini 3 Pro, described as its “most factual” model, specifically trained to minimize hallucinations in complex, long-running agentic tasks where multiple autonomous decisions are required.
To demonstrate performance gains, Google introduced and open-sourced a new benchmark called DeepSearchQA for evaluating agents on multi-step information-seeking tasks. The company also tested the agent on independent benchmarks including Humanity’s Last Exam and BrowserComp.
Google’s agent topped both its own DeepSearchQA benchmark and Humanity’s Last Exam, while OpenAI’s ChatGPT 5 Pro finished a close second overall and slightly edged out Google on BrowserComp.
The comparisons were rendered outdated almost immediately, as OpenAI released its anticipated GPT 5.2 (codename Garlic) on the same day, with claims of superior performance across various standard benchmarks.
The simultaneous timing of the announcements appeared deliberate, with Google releasing major AI news on the same day the industry was focused on OpenAI’s Garlic launch.
Google introduces Gemini in its Pixel Watch 4
Google announced some significant changes to the Pixel Watch 4 this week. New AI-powered gestures are launched in the latest smartwatch.
This upgraded version improves the convenience of the users to a great extent. One hand is all it takes to control the user’s watch.
The new gestures are known as the two-pinch and Wrist turn gestures.
As a result, the tasks are handled without having to touch the screen by the users. It is a hands-free application that is suitable for busy users.
These particular motions enable one to roll through notifications. Snoozing an alarm is also easily done.
Moreover, the gestures control the playback of music and timers. They even allow users to respond to and dismiss calls. On-screen notifications can also be displayed on the watch.
Also read: Best Gemini AI Photo Prompt Copy & Paste Collections
