Multimodal
Reason over images, PDFs, audio, and video together.
Reason across images, PDFs, and more
Multimodal is where Gemini genuinely stands out: it reads images, PDFs, charts, diagrams, and other formats and reasons over them alongside text in a single prompt. For research, that means you can hand it the evidence in its original form — a chart, a scanned document, a screenshot — instead of transcribing it into words and losing detail. Combine that with its long context and you can pose a question that spans several documents and images at once, and get an answer grounded in all of them.
Show it, don't retype it
The Gemini-specific research habit is to show rather than describe. Rather than summarising a chart or paraphrasing a page, give Gemini the chart and the page and let it read them directly. This preserves the detail that matters and lets it reason over the actual source. Pair that with a sharp, decision-focused question — what you are trying to work out and why — and Gemini turns a pile of mixed-format material into an answer you can act on.
What does multimodal mean in Gemini?
Gemini can read and reason over images, PDFs, charts, diagrams, and other formats alongside text in one prompt — so you can give it evidence in its original form rather than transcribing it.
How do I research with images and documents in Gemini?
Show it the material directly — charts, screenshots, PDFs — rather than describing it, and pair that with a precise, decision-focused question. Its long context lets it span several sources at once.
Why show Gemini a chart instead of describing it?
Because transcribing loses detail. Giving Gemini the chart itself lets it read the actual data and reason over the real source, which produces more accurate analysis.