Raycaster / evalsTask overviewSpreadsheetBenchCategory

SpreadsheetBench · Gemini 3.1 Pro · dual

37456

1/1Pass

Gemini 3.1 Pro on SpreadsheetBench: 37456 (dual harness). Browse score, rubric, and public trace.

1/1 · Pass
SpreadsheetBench
AI Agents for Spreadsheet Automation and Workbook Editing

Prompt excerpt

Task context

You are solving a spreadsheet benchmark task in a real workbook. Objective: Produce the correct final workbook state for the expected answer region. What matters: - Only the values in the expected answer region will be graded. - The workbook is the answer. Instructions: 1. Read the workbook and inspect the relevant data region first. 2. Infer the required result for the provided workbook instance. 3. Write the final value(s) directly into the expected answer region. 4. Do not rely on prose, formulas in your chat response, pseudocode, or VBA as the answer unless the benchmark explicitly requires those to be written into cells. 5. If the natural-language task asks for a general method, formula, or macro, convert that into the concrete result needed for this workbook instance. 6. Keep your final text response short and only summarize the workbook cells you changed. Relevant data region(s): Expected answer region(s): G12:J15 Expected answer sheet(s): Task: I need to calculate daily product sales using a custom 24-hour window that runs from 6:00 AM of the current date to 5:59:59 AM of the next date. The formula must: Dynamically reference the product from the header so it can be copied across columns without editing. Use the next row’s date as the end boundary for each window. Calculate each day’s sales using a cumulative subtraction method (sum through the end of the window minus sales beyond it). Only run when a next date exists—if the current row is the last date in the dataset, the formula should not calculate. Once done change the formatting of cells G12:J15 from accounting to currency.

Response trace

Agent response, tools, files, and edits

Open full trace

On mobile, open the full trace for the best view of the agent response, tool calls, file reads, edits, and grading context.

Open mobile trace