Google is broadening the reach of its Gemini 2.5 Pro Experimental artificial intelligence (AI) model by integrating it into additional tools. On Tuesday, the tech powerhouse from Mountain View announced that its latest Gemini model will now enhance the agentic Deep Research tool. This feature remains exclusive to the platform’s paid subscribers, while free users will continue utilizing the AI agent with the experimental Gemini 2.0 Flash Thinking model. The company highlights that users can expect significant advancements in analytical reasoning capabilities with the new large language model (LLM).
Gemini 2.5 Pro Will Now Power Deep Research
In a blog post, Google announced the incorporation of its latest foundation model into Deep Research. The Gemini 2.5 Pro Experimental AI model, unveiled last month, comes with numerous enhancements. This model is notable for being the first in the Gemini series to include integrated reasoning capabilities, meaning that all Gemini 2.5 AI models will operate as native “Thinking” models, facilitating a smooth integration into Deep Research.
According to Google, an internal evaluation of Deep Research using the Gemini 2.5 Pro Experimental showed that raters preferred the resulting reports over those produced by competitor tools. Testers particularly noted significant improvements in analytical reasoning, the synthesis of information, and the creation of insightful research reports.
Presently, access to Deep Research powered by Gemini 2.5 Pro Experimental is restricted to Gemini Advanced subscribers. These users can navigate the AI agent through web interfaces and mobile applications on both Android and iOS platforms. Those on the free tier of the Gemini platform will still have the option to use Deep Research with the Gemini 2.0 Flash Thinking model.
It’s worth noting that Google first launched the agentic Deep Research feature in December 2024, initially available only to Gemini Advanced subscribers. In 2025, the AI agent was made accessible to a broader audience.
Deep Research is capable of creating comprehensive multi-step research plans, performing web searches, and gathering information on various topics and related fields. It subsequently analyzes the data collected to generate a detailed report, presenting the findings to the user.
