What it measures
Spreadsheet and document questions test the quantitative work expected of an analyst.
How to read the result
Look for correct joins, units, denominators, and evidence supporting a numerical answer.
Limits and comparison traps
Access to code and spreadsheet tooling is part of the configuration. A plausible number is weaker evidence than an auditable computation.
Benchmark reference notes
Private data-analyst benchmark requiring spreadsheet and document manipulation to answer quantitative business questions.
Where this fits in an evaluation suite
A transcript that says “done” is weaker evidence than the verified final state. Track policy violations, recoveries, and invalid attempts alongside success.
Start with a capability hypothesis, then choose a second test with a different failure mode. Keep the checkpoint, reasoning effort, agent tools, sampling policy, and scoring procedure fixed when measuring a training intervention. Record failures and missing outputs in the denominator.
A reproducible comparison
Use the exact implementation, version, task subset, tool access, attempt budget, and judge described by the source.
Publish the exact run configuration, item coverage, score units, and available uncertainty. If a source supplies only an aggregate or a single run, preserve that limit. A new model name or a rounded score alone does not establish a capability improvement.
PRIMARY EVIDENCE & FURTHER READING
Reviewed 2026-10-26. Explanation and interpretation by RLTasks. Benchmark ownership remains with its original authors.