OfficeQA Pro
A benchmark designed to evaluate the comprehensive capabilities of large language models in office document question-answering scenarios. It covers complex tasks such as table comprehension, cross-page reference resolution, and mixed text-image parsing across Word, Excel, and PowerPoint formats, closely reflecting real enterprise use cases.
Core Facts
GPT-6 Sol and GPT-6 Luna deliver strong performance at lower cost and have pushed the Pareto frontier on the OfficeQA Pro benchmark
The OfficeQA series of benchmarks focuses on question-answering tasks in office scenarios, evaluating a model's ability to understand and retrieve content from enterprise documents, spreadsheets, reports, and similar materials.
Timeline (last 90 days)
Nace AI said Drex for Search reduced the Agent's interaction turns on OfficeQA Pro, but did not disclose the specific magnitude of the reduction.
GLM-5.3-Flash scores 78.4 on Toolathlon Verified and 62.4 on OfficeQA Pro, both ahead of Claude Opus 4.8.
GPT-6 Sol and GPT-6 Luna deliver strong performance at lower cost and have pushed the Pareto frontier on the OfficeQA Pro benchmark
Benchmark data cited by vendors or platform providers are typically best-case results obtained under controlled conditions, and enterprises need to conduct independent evaluations based on their own scenarios
The OfficeQA series of benchmarks focuses on question-answering tasks in office scenarios, evaluating a model's ability to understand and retrieve content from enterprise documents, spreadsheets, reports, and similar materials.
According to Databricks' internal evaluation, GPT-6 Astra has reached industry-leading levels on enterprise-grade benchmarks including OfficeQA Pro, OfficeQA Pro V2, and Document Processing.
OfficeQA Pro is a benchmark designed specifically to evaluate models' comprehensive capabilities in office document question-answering scenarios, covering tasks such as table comprehension, cross-page references, and mixed text-image parsing in Word, Excel, PowerPoint, and other formats.
All Facts (6)
GPT-6 Sol and GPT-6 Luna deliver strong performance at lower cost and have pushed the Pareto frontier on the OfficeQA Pro benchmark
75%VerifiedThe OfficeQA series of benchmarks focuses on question-answering tasks in office scenarios, evaluating a model's ability to understand and retrieve content from enterprise documents, spreadsheets, reports, and similar materials.
75%UnverifiedAccording to Databricks' internal evaluation, GPT-6 Astra has reached industry-leading levels on enterprise-grade benchmarks including OfficeQA Pro, OfficeQA Pro V2, and Document Processing.
70%UnverifiedNace AI said Drex for Search reduced the Agent's interaction turns on OfficeQA Pro, but did not disclose the specific magnitude of the reduction.
50%UnverifiedBenchmark data cited by vendors or platform providers are typically best-case results obtained under controlled conditions, and enterprises need to conduct independent evaluations based on their own scenarios
50%UnverifiedOfficeQA Pro is a benchmark designed specifically to evaluate models' comprehensive capabilities in office document question-answering scenarios, covering tasks such as table comprehension, cross-page references, and mixed text-image parsing in Word, Excel, PowerPoint, and other formats.
50%