Gemini 2.5 Pro Experimental Will Now Power the Agentic Deep Research Feature

Google is expanding its Gemini 2.5 Pro Experimental artificial intelligence (AI) model to other tools. On Tuesday, the Mountain View-based tech giant announced that its latest Gemini model will now power the agentic Deep Research tool. This feature is currently limited to the paid subscribers of the AI platform, while the free users will continue to use the AI agent with the Gemini 2.0 Flash Thinking (experimental) model. Google says the latest large language model (LLM) will allow users to see noticeable improvements in the tool's analytical reasoning capability.

Gemini 2.5 Pro Will Now Power Deep Research

In a blog post, the tech giant announced the expansion of its latest foundation model to Deep Research. The Gemini 2.5 Pro Experimental AI model was introduced last month with several improvements. This is also the first Gemini series to feature integrated reasoning capability, which means all the Gemini 2.5 family AI models will be natively “Thinking” models. Because of this, integrating the model into Deep Research becomes a seamless process.

Google claimed that it conducted an internal test of Deep Research with Gemini 2.5 Pro Experimental, and the raters preferred the generated reports over competitors. In particular, the company claimed that the testers noticed a noticeable improvement in analytical reasoning, information synthesis, and generating insightful research reports.

Currently, Deep Research with Gemini 2.5 Pro Experimental is only available to Gemini Advanced subscribers. Users with the subscription can access the AI agent across the web as well as Android and iOS apps. Meanwhile, those on the free tier of Gemini can continue to use Deep Research with the Gemini 2.0 Flash Thinking (experimental) model.

Notably, Google first released its agentic Deep Research in December 2024. However, initially, it was only available to Gemini Advanced users. Later, in 2025, the AI agent was expanded to all users.

Deep Research can create multi-step research plans, run web searches, and collect information on the topic and related fields. After that, it analyses the gathered data, creates a detailed report and shows the output to the user by preparing a detailed report.