SpreadsheetBench · GPT-5.4 · dual
57113
GPT-5.4 on SpreadsheetBench: 57113 (dual harness). Browse score, rubric, and public trace.
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): B2:C3 Expected answer sheet(s): Task: I am working with Twitter data and attempting to create a new Excel column that extracts all hashtags from each tweet. While I can successfully use a formula to pull the first hashtag from a cell, I am struggling to modify it or use another method to extract multiple hashtags from cells that contain more than one. Additionally, I want to output these hashtags in adjacent cells starting from column B, without the # symbol and without any changes to the header row. An example file has been provided. Modify my current formula to achieve this. My current formula is: TRIM(LEFT(SUBSTITUTE(MID(A2,FIND("#",A2),LEN(A2))," ",REPT(" ",100)),100))
Response trace
Agent response, tools, files, and edits
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