OfficeQA Pro evaluates how models handle parsing, retrieval, and grounded reasoning across workflows involving scanned PDFs, legacy files, and long-context documents, tasks that frequently break production agent systems.

In the agent-harness setting, GPT‑5.5 reduced errors by 46% compared to GPT‑5.4 and became the first model to surpass 50% accuracy on OfficeQA Pro.

OfficeQA Pro includes large numbers of scanned or legacy enterprise documents, where small extraction errors during parsing can cascade downstream through the rest of the workflow. “Once you can’t extract a certain digit or number, that changes the entire trajectory of what the agent works with,” Singhvi explains.

Databricks saw the largest gains from GPT‑5.5 in these parsing-heavy workflows. “Earlier models like 5.4 were unable to parse all the digits correctly, but it seems like 5.5 has a step-function lift in parsing older documents and scanned PDFs,” Singhvi says.