reasoning benchmark
OfficeQA Pro
OfficeQA Pro evaluates AI models on professional knowledge-work questions and tasks drawn from real office workflows, including document analysis, spreadsheet reasoning, and information synthesis across business domains.
Updated Aug 11, 2026
Models7
Model coverage7
MetricScore
EvidenceB
OfficeQA Pro Ranking
Higher score ranks better on this benchmark.
| 01 | BY | 72.2% | 100.0% | 7 | C | |
| 02 | BY | 71.1% | 83.3% | 7 | C | |
| 03 | AN | 66.2% | 66.7% | 7 | C | |
| 04 | MA | 63.3% | 50.0% | 7 | C | |
| 05 | AN | 59.4% | 33.3% | 7 | C | |
| 06 | OP | 54.1% | 16.7% | 7 | C | |
| 07 | MI | 45.1% | 0.0% | 7 | C |
OfficeQA Pro Score Distribution
A closer view of the leading scores on this benchmark.
OfficeQA Pro Highlights
The leading models and scores on this benchmark.
What is OfficeQA Pro?
What OfficeQA Pro measures and how its scores work.
OfficeQA Pro evaluates AI models on professional knowledge-work questions and tasks drawn from real office workflows, including document analysis, spreadsheet reasoning, and information synthesis across business domains.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- OfficeQA Pro
- Modality
- text
- Primary category
- reasoning
- Score direction
- higher
- LLMBoard eligible
- Yes
- Evaluation key
- officeqa-pro|llm-stats-current
Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about OfficeQA Pro.
Which model scores highest on OfficeQA Pro?
Seed 2.1 Pro is currently ranked first with 72.2%.
What does OfficeQA Pro measure?
OfficeQA Pro evaluates AI models on professional knowledge-work questions and tasks drawn from real office workflows, including document analysis, spreadsheet reasoning, and information synthesis across business domains.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
7 model results are currently shown.
Does this benchmark affect the overall score?
Yes. This benchmark can contribute to the current LLMBoard capability score.