SpreadsheetBench
22-47
SpreadsheetBench task 22-47. Inspect the exact spreadsheet prompt, compare published model runs, open the agent response trace, and review grades for workbook editing.
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): sheet1!A1:j24',',ورق1!B1:B11' Expected answer region(s): F2:H10 Expected answer sheet(s): sheet1 Task: I need to sort data in column B based on a helper column (J) so that the first entries written in column J are shown first in the output range F:H. Names listed in column J, include all matching rows and keep their original order from the source and do not sort within the group. Names not listed in column J should be arranged as they are in the original data (same order). The sort should skip empty cells, headers, and duplicate items, where duplicates are defined by identical entries in both column B and C. If the helper column J is empty, only then sort alphabetically A–Z and still skip empty cells, headers, and duplicate items. Additionally, I need this sorting process to handle new ranges that contain headers and some names in range A:C, as well as changes in names listed in column J. The final answer should be output in columns G and H, and sort only column H sorted lowest to highest.
Published trajectories
Agent runs on this task
Curated dual-harness runs. Best scored run per model.
| Model | Harness | Score | Result | Links |
|---|---|---|---|---|
| GPT-5.4showcase | dual | 0/1 | Fail | Share pagePublic trace |
| Gemini 3.1 Pro | dual | 0/1 | Fail | Share pagePublic trace |