Raycaster / evalsBack to SpreadsheetBench

SpreadsheetBench

51090

1/1Pass

SpreadsheetBench task 51090. Inspect the exact spreadsheet prompt, compare published model runs, open the agent response trace, and review grades for workbook editing.

Spreadsheet editingDual harness
ssb-51090
SpreadsheetBench
1 model
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): Expected answer region(s): Daily Numbers'!Q3:Q24 Expected answer sheet(s): Task: How can I calculate the difference between the total cartons received as indicated in Column M and the errors present in Columns N to R, while filtering for a specific warehouse (e.g., warehouse 27), error codes (specifically for each of II, IR, IT, OV, PI with code II in column N), date, and users (e.g., CHROGIL1, CHDSPOLJ, CHSJEFFE, CHBTHOMA for each error code) across a data set? Additionally, for the IR error code listed in Column O, I only want to include positive numbers from the Errors Table. The goal is to simplify a formula that can achieve this, as my current one is too complex and not providing the correct value, such as the expected 699 for Q2.

Published trajectories

Agent runs on this task

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

ModelHarnessScoreResultLinks
GPT-5.4showcasedual1/1Pass