Powered by

Home Trending News Google unveils upgraded Gemini Deep Research Agent, just hours before OpenAI launches GPT-5.2

Google unveils upgraded Gemini Deep Research Agent, just hours before OpenAI launches GPT-5.2

Google rolls out upgraded Gemini Deep Research, remodelling what was once a niche research-report generator into a full-scale autonomous agent.

By Ishita Ganguly
New Update
Sam Altman

Sam Altman (left); Sundar Pichai (right)

Listen to this article
0.75x1x1.5x
00:00/ 00:00

Google has introduced a sweeping upgrade to Gemini Deep Research, modifying what was once a niche research-report generator into a full-scale autonomous agent built for long-form reasoning and enterprise-grade analysis. The announcement arrived impeccably on the very same day OpenAI launched GPT-5.2, its latest flagship model.

Advertisment

Google rolls out Gemini Deep Research

Powered by Gemini 3 Pro, Google’s most advanced and most “factual” system to date, the revamped Deep Research agent is engineered to minimise hallucinations during operations that may run for hours or days.

In such high-stakes workflows, a single flawed inference can cascade into major downstream errors, a risk Google says it has aggressively targeted.

A centrepiece of the rollout is Google’s new Interactions API, allowing developers to embed deep-reasoning agents directly into their own products.

Instead of producing static reports, the upgraded agent can be connected to corporate tools to handle due diligence, drug safety reviews or any process requiring the synthesis of vast, context-heavy information.

Google is also weaving Deep Research into its consumer ecosystem, from Search and Finance to the Gemini app and NotebookLM.

It is a clear signal that the company envisions an era where users no longer search manually but delegate entire investigative tasks to AI.

To validate its claims, Google introduced a benchmark called DeepSearchQA, measuring multi-step information-seeking performance.

On both that test and the external Humanity’s Last Exam benchmark, Gemini Deep Research topped the charts.

Results from BrowserComp, however, showed OpenAI’s ChatGPT 5 Pro edging ahead.

The rivalry intensified within hours as OpenAI’s freshly released GPT-5.2, codenamed “Garlic,” claimed widespread gains in reasoning and accuracy.

The synchronised releases clearly point to a rapidly accelerating contest for dominance in agentic AI.

Also read: Mamaearth parent Honasa Consumer acquires Reginald Men in ₹195 Cr deal (startuppedia.in)