Deep Research Gemini Now Supported By 2.5 Pro Experimental Model

Gemini present the 2.5 Pro model (photo: dock. unsplash)

JAKARTA Google has launched the Deep Research feature to all Gemini users in mid-March 2025, but with the support of 2.0 Flash Thinking. Now, the feature is supported by the 2.5 Pro model.

According to a 9to5google report, the 2.5 Pro model is only available to Gemini Advanced users and is not available to free customers. However, based on VOI findings, the experimental model can also be found on a free account.

Deep Research is a feature provided to assist in-depth research. In fact, Google calls this feature a personal Artificial Intelligence (AI) research assistant. With the support of 2.5 Pros, Google guarantees smarter analysis results.

To prove the ability of Deep Research 2.5 Pro, Google conducted joint testing with a number of appraisers. When compared to Deep Research from OpenAI, Google stated that research providers prefer their models.

"In our tests, appraisers prefer reports generated by Gemini Deep Research backed by the 2.5 Pro over other leading in-depth research providers with margins of more than 2 to 1," Google said in a statement.

In a survey conducted by Google on researchers, about 69.9 percent of researchers prefer Deep Research 2.5 Pro. The reason is, the research model is easier to follow instructions, more complete in reaching information and topics, and the quality of writing is considered much more concise and clear.

"Users testing Deep Research on our latest models report a marked increase in analytical reasoning, information synthesis, and produce more in-depth research reports," Google explained.

If you want to test a contested experimental model, you can access the 2.5 Pro using Gemini versions of the web, Android, or iOS. The model is available without restrictions for Gemini Advanced users.