Gemini in Google Sheets Achieves State-of-the-Art on SpreadsheetBench

Google announced that Gemini in Sheets has reached state-of-the-art performance on SpreadsheetBench, a public benchmark that evaluates AI models on their ability to edit spreadsheets in real-world scenarios.

The model achieved a 70.48% success rate, which Google says "not only exceeds competitors but nears human expert ability."

The update adds new beta features for Gemini in Sheets that let users "create, organize and edit entire sheets, from basic tasks to complex data analysis—just describe what you need."

SpreadsheetBench is a public benchmark designed to test real-world spreadsheet manipulation. According to its documentation, the benchmark evaluates models on their ability to understand spreadsheet structures, perform cell edits, and produce correct outputs across a range of practical scenarios.

The announcement is part of Google's broader push to embed AI assistants directly into productivity tools. Gemini's strong performance on spreadsheet tasks positions it ahead of competing models in a domain that requires understanding both structured data and natural language instructions.