Google Debuts Enhanced AI Research Tool as OpenAI Introduces GPT-5.2
In a significant development in artificial intelligence, Google has launched a revamped version of its research agent, Gemini Deep Research, built on the cutting-edge Gemini 3 Pro foundation model. This innovative agent goes beyond merely generating research reports; it also empowers developers to integrate Google’s advanced SATA-model research capabilities directly into their applications. This functionality is enabled by the new Interactions API, aimed at providing developers with greater control in the evolving landscape of agent-based AI.
The Gemini Deep Research tool is designed to synthesize vast amounts of information and efficiently process complex prompts. Google highlights its utility across various fields, including due diligence and drug safety research. Furthermore, the company plans to incorporate this advanced research agent into key services such as Google Search, Google Finance, and its Gemini App, shifting towards a future where AI agents, rather than users, perform online searches.
Gemini Deep Research benefits from the Gemini 3 Pro model, touted as Google’s “most factual” AI, which is specifically trained to reduce the occurrence of hallucinations during intricate tasks. AI hallucinations, where models generate incorrect or fabricated information, pose a significant risk in long-duration decision-making processes. A single erroneous choice can jeopardize the entire outcome, making reliability critical.
To substantiate its advancements, Google has introduced a new performance benchmark called DeepSearchQA, designed to evaluate agents on complex, multi-step information retrieval tasks. This benchmark has been made open source. Additionally, Google tested Deep Research using independent benchmarks, notably Humanity’s Last Exam and BrowserComp, resulting in strong performance against competitors. While Google’s new agent excelled in its own benchmarks, OpenAI’s ChatGPT 5 Pro closely trailed and outperformed Google in BrowserComp.
The timing of Google’s announcement coincides with OpenAI’s launch of the highly anticipated GPT-5.2, codenamed Garlic. OpenAI claims that this latest iteration surpasses its competitors, including Google, across several traditional benchmarks, adding a competitive edge to the ongoing AI race.
Key Highlights:
– Google introduces Gemini Deep Research built on the Gemini 3 Pro model.
– Developers can now integrate advanced research capabilities into their applications via the Interactions API.
– The tool addresses complex research tasks like due diligence and drug safety.
– Gemini Deep Research aims to minimize AI hallucinations, enhancing reliability.
– Google unveils a new open-source benchmark, DeepSearchQA.
– OpenAI launches its GPT-5.2 model on the same day, showcasing competitive advancements.
