Raycaster / evalsBack to SpreadsheetBench

SpreadsheetBench

280-17

1/1Pass

SpreadsheetBench task 280-17. Inspect the exact spreadsheet prompt, compare published model runs, open the agent response trace, and review grades for workbook editing.

Spreadsheet editingDual harness
ssb-280-17
SpreadsheetBench
2 models
Dual (parsed + original)
PromptWorkbook objective and answer cells
ResponseOpen each trace to inspect tool use and edits
GradeCompare score and pass/fail by model

Task prompt

What the agent was asked to do

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): A1:B12 Expected answer region(s): A1:B12 Expected answer sheet(s): Sheet1 Task: How can I remove all rows from an Excel spreadsheet except those with the last unique entry in a specific column? I have a file with 1100 rows and 36 columns, and within one column there are some numbers that are duplicates. Specifically, in my example file, I want to keep the rows that contain the last unique number in the column, such as row 3, row 6, and so on, and then delete the other rows with duplicate numbers. An example file is provided for your reference.

Published trajectories

Agent runs on this task

Curated dual-harness runs. Best scored run per model.

ModelHarnessScoreResultLinks
GPT-5.4showcasedual1/1Pass
Gemini 3.1 Produal1/1Pass