Business Bench

Tasks · desk · Reports

August yield and scrap by production line

production-yield · timeout 1200 s · source

The ask

Sent verbatim as the prompt. The agent works in a copy of the workspace below and sees nothing else.

Tomasz wants August's yield and scrap for each production line from the batch records. Save it as yield.xlsx with live formulas; his note in the folder says how he wants it worked out.

Workspace

  • batch_records_2026-08.csv
  • note_from_tomasz.txt
  • product_specs.xlsx

Expected deliverables

  • yield.xlsx

This task also has a task-specific grader (check.py).

Checks

The task passes only when every required check passes. Reference answers, tolerances, and the planted traps are in the repository, not here.

CheckTypeRequiredFailed (frozen) / 6
yield.xlsx existsfile_existsyes0
live formulasxlsx_has_formulasyes0
no error cellsxlsx_no_errorsyes3
Bar line good output kg (cases and pieces converted)xlsx_value_presentyes3
Bar line scrap kg (rework netted)xlsx_value_presentyes3
Granola line good output kg (bags converted)xlsx_value_presentyes3
Bites line scrap kg (grams converted, rework netted)xlsx_value_presentyes3
plant input kg (no rework or trial input)xlsx_value_presentyes3
yield per linecustomyes3

Recorded results, complete desk comparison

pass pass, frozen scorer only fail timed out
SystemRepetitionsFrozen / 3Raw / 3Median minutesMean cost (USD)
Proto / DeepSeek V4.1 Flash337.00.0704
Codex / GPT-5.6 sol006.00.6744

Per-attempt check verdicts, raw and frozen verdicts, usage, receipt and artifact hashes for this task are in results/latest/attempts.jsonl under "task": "production-yield".