Google has released the Gemini Deep Research agent through the Interactions API, making their advanced autonomous research technology accessible for developers to embed into their own applications. This move opens direct access to Google’s latest reasoning core, Gemini 3 Pro, for a range of industries. The feature targets developers, financial analysts, biotech researchers, and market intelligence teams who require detailed, multi-step research and synthesis across vast information landscapes. The release is available to the public for integration via the Gemini API key in Google AI Studio, and will soon roll out across Google Search, NotebookLM, Google Finance, and the Gemini App, with enterprise plans for Vertex AI.
Say hello to the new Interactions API and our first agent, Gemini Deep Research, now available for developers 🤖!
— Logan Kilpatrick (@OfficialLoganK) December 11, 2025
The Interactions API is a new unified interface to interact with both models and agents. Our Deep Research agent is also SOTA on many dimensions...
Gemini Deep Research operates by autonomously gathering, analyzing, and synthesizing data from the web and user-uploaded documents (PDFs, CSVs, docs). It iteratively plans its research, identifies knowledge gaps, and navigates deeply into websites, supporting long-running, complex research tasks. The agent is optimized for factual accuracy, featuring granular sourcing and citation for claims, report steerability via prompts, and structured outputs in JSON for downstream processing. Benchmarks show it achieves 46.4% on Humanity’s Last Exam, 66.1% on DeepSearchQA, and 59.2% on BrowseComp, setting new standards for comprehensive research.
Google is positioning this release to support research-heavy fields. Early adopters in finance report increased efficiency in due diligence, while biotech companies like Axiom Bio note deeper initial insights for drug discovery. Developers gain a unified tool for information synthesis, detailed reporting, and easy integration, distinguishing Gemini Deep Research from previous offerings and rival agents in both depth and flexibility.
