I've been testing Gemini deep research upgraded with 2.5 Pro this morning.

Honestly, it's a great upgrade over the 2.0 Flash Thinking version, which makes it often better than OpenAI's Deep search in terms of information found.

It also generates tables and appendixes based on the aggregated search data, which is quite amazing (and they're exportable to Google Sheets)

HOWEVER

The tone is still very academic and dry, and the paragraphs are overly long, even if you prompt for something more engaging.

Still, it's an INCREDIBLE research tool and context builder for LLMs.

This is especially true when you consider that you get unlimited Deep Research with Google Workspace or Google One vs. just 10/month with ChatGPT.


This post was originally shared by on Linkedin.