Google introduces a new version of the Gemini Deep Research research tool. This tool is built based on the Gemini 3 Pro model and is designed to process large amounts of data, synthesize in-depth information and support complex tasks such as appraisal or assessing the safety of drugs.
Google says Deep Research does more than just create automatic reports. Users can integrate this research capability into their own application through the new Interactive API. This is a preparation for the AI assistant phase that plays a central role in searching and processing information.
In the coming time, Deep Research will be included in many of the company's services including Google Search, Google Finance, Gemini application and NotebookLM. Google affirmed that the Gemini 3 Pro model helps reduce AI hallucinations in tasks that require multi-step reasoning.
Google also announced the DeepSearchQA review standard to test its ability to search for and process complex information. This tool is being tested on other independent testing sets. The results showed that Deep Research scored high. However, OpenAI's ChatGPT 5 Pro is still close and has better results in some browser-based reviews.
This development comes as OpenAI announces the GPT 5.2 model. The new model is introduced to have outstanding performance in many tests and is considered a direct response to Google's advanced statements on the same day.
